Available online at www.sciencedirect.

com

Talanta 74 (2008) 14631475

Development and optimisation of a single extraction procedure

for the LC/MS/MS analysis of two pharmaceutical
classes residues in sewage treatment plant
Anne Piram, Arnaud Salvador , Jean-Yves Gauvrit, Pierre Lanteri, Rene Faure
Universite de Lyon, Lyon 1, UMR 5180, CNRS, Laboratoire des Sciences Analytiques, 43 Boulevard du 11 novembre 1918,
69622 Villeurbanne, France
Received 27 April 2007; received in revised form 5 September 2007; accepted 28 September 2007
Available online 9 October 2007

Abstract
A unique extraction procedure leading to the separation of 2 different pharmaceutical classes molecules has been developed and optimised
by chemometric tools. From only one sampling, this analytical method allows the determination of 21 pharmaceuticals from corticosteroids
and -blockers classes. Performing the SPE on Oasis MCX (mixed-mode cation exchange), the sequential elution of each pharmaceutical class
is achievable, allowing a high purity level of extracts as well as high recovery rates. Performing a unique sample preparation results in an
important save of time. The extracts were then analysed by LC/MS/MS, using a Hibar Purospher Star column for -blockers and an X-Bridge
column for corticosteroids with formate buffer (pH 3.8)/AcN and water/AcN mobile phases, respectively. This work also includes a study of the
chromatographic and mass spectrometric parameters in order to increase the analyte signal. The optimised SPE-LC/MS/MS method was then
applied to environmental samples from sewage treatment plant (STP). -Blockers and corticosteroids were detected, respectively, in concentrations
up to 318 ng L1 (sotalol) and 174 ng L1 (cortisone), in STP influents. Moreover, both pharmaceutical classes have also been detected in STP
effluents. As far as we know, this is the first paper reporting the detection of corticosteroids in environmental waters. The developed analytical
method can be used in further studies to investigate the environmental contamination by these drugs.
2007 Elsevier B.V. All rights reserved.
Keywords: Pharmaceuticals; Water analysis; SPE; -Blockers; Steroids; Liquid chromatography; Mass spectrometry

1. Introduction
Quantification of drugs in the environment is a matter of concern since they potentially represent an important environmental
risk [14]. Indeed, these substances have been developed in
order to perform a biological effect and are partially excreted
unchanged by the body. Thus, they can affect non-targeted
organisms and be harmful for the ecosystem.
Human medications reach the environment through sewage
treatment plant (STP). Effectively it was recognised that STP
efficiency is very low for this kind of micropollutants, allowing the passage of these compounds and their release into the
environment. Several authors have reported the environmental occurrence of drugs in surface-, ground- and tap-water,

Corresponding author. Tel.: +33 4 72 43 11 52; fax: +33 4 72 44 62 02.

E-mail address: arnaud.salvador@univ-lyon1.fr (A. Salvador).

0039-9140/$ see front matter 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.talanta.2007.09.038

as well as in STP influents and effluents [512]. Many studies report the quantification of large amounts of non-steroidal
anti-inflammatory drugs (aspirin, paracetamol) [1315], antidepressant (fluoxetine, paroxetine) [15], antibiotics (amoxicillin,
sulfamethoxazole) [1517], steroid sex hormones (estrone,
estriol) [1820] or cardiovascular drugs (atenolol, dilitiazem)
[11,15,16,21]. -Blockers (cardiovascular drugs) have been
reported to contaminate both American and European waters
in a large number of studies [13,2125], with concentrations
up to 2000 ng L1 . The occurrence of steroids in environmental
waters has been confined to the study of the estrogens subclass
[1820]. Few papers report the contamination by other steroid
class molecules, such as cholestanes (cholesterol), progestagens (progesterone, levonorgestrel) or androgens (testosterone)
[2628]. However, no environmental measurements were done
concerning corticosteroids although they are consumed in large
quantities as anti-inflammatory or antiallergenic drugs. Moreover, this pharmaceutical class has been underlined as a potential

1464

A. Piram et al. / Talanta 74 (2008) 14631475

environmental risk by a study of prioritisation based on a modelisation [29].

The aim of this study was to analyse simultaneously 2
pharmaceutical classes: -blockers and corticosteroids whose
structures are given in Tables 1 and 2 , respectively. These
pharmaceutical classes were selected because of their high consumption (e.g. 35 t year1 in France for propranolol -blocker)
and the information deficiency for the corticosteroids impact
in the environment. Analytical method development for corticosteroid compounds analysis in environmental samples was
necessary since so far no methods were proposed by the literature
concerning this matrix (water).
The difficulty of tracking trace pollution of drugs in STP is
linked with the complexity of environmental matrices. Indeed,
they contain lots of other organic pollutants such as pesticides or
personal care products. Moreover, there is an important dilution
factor in the environment which results in expected environmental concentration in the ng L1 range, as it has been found in both
European and American waters [528]. Thus, it is necessary to
concentrate and purify samples before analysis. In the literature,
this sample preparation is mainly done by solid phase extraction
(SPE) [528,3033]. The development of multiresidue methods
is important in environmental analysis of pharmaceuticals and
continuous efforts have been made to design new analytical procedures able to quantify traces amounts of pharmaceuticals in
environmental waters. These recently developed methods aim
to improve the total analysis time and reduce the analysis cost.
Then, it is useful to develop a unique extraction procedure, which
allows the extraction of both pharmaceutical classes from only
one sampling, on the same SPE sorbent, with a sequential final
elution of each class.
The extracted sample analysis should be then carried out by
means of sensitive and selective detection tools, such as the
tandem mass spectrometers, which are widely used in environmental analysis. A liquid chromatographic device is coupled to
the mass spectrometer as reported in literature [528,3033].
This paper also presents a study for the LC/MS/MS parameters
optimisation in order to improve the analyte response.
2. Experimental
2.1. Chemicals
Betamethasone (97.8%) (Be), budesonide (99.2%) (Bu), cortisol (98%) (Hy), cortisone (98.3%) (C), dexamethasone (98%)
(D), flunisolide (98.6%) (Fs), fluocinolone acetonide (99.6%)
(Fc), prednisolone (99%) (Pl), prednisone (99%) (P), triamcinolone (98.2%) (T), triamcinolone acetonide (99.3%) (TA),
acebutolol hydrochloride (100%) (Ac), alprenolol hydrochloride (99%) (Al), atenolol (99%) (At), metoprolol tartrate salt
(100%) (M), nadolol (100%) (N), propranolol hydrochloride
(99%) (Pr), pindolol (99%) (Pi), sotalol hydrochloride (99%)
(S), and timolol maleate salt (99.9%) (Ti) were provided by
SigmaAldrich (Saint Quentin Fallavier, France). Bisoprolol
fumarate (100%) (Bi) was provided by Sequoia research (Pangbourne, United Kingdom). Methanol HPLC grade (MeOH)
was provided by SDS-Carlo-Erba (Peypin, France). Acetonitrile

LC/MS grade (AcN), formic acid (85%) (FOA) and ammonium formate (99%) were obtained from Fisher Scientific (Val
de Rueil, France). Ammonia solution RP Normapur (28%) was
provided by Prolabo (Fontenay sous Bois, France). Ultra-pure
water was generated by an Elgastat UHQ PS from Elga LabWater
(High Wycombe Bucks, United Kingdom).
Each corticosteroid stock solution was prepared in AcN at
1000 mg L1 , excepted for triamcinolone and triamcinolone
acetonide (250 mg L1 ) which have a lower solubility. Each blocker stock solution was prepared in MeOH at 1000 mg L1 .
A first 50 mg L1 standard working solution (WSc ) containing
the 11 corticosteroids was prepared in AcN. A second 50 mg L1
standard working solution (WS) containing the 10 -blockers
was prepared in MeOH. These solutions were stored at 4 C, in
the dark.
4, 2 and 1 mg L1 working solutions were prepared
by dilution of WSc in water/AcN/FOA (75:25:0.1, v/v/v).
Corticosteroids calibration standards were then prepared in
water/AcN/FOA (75:25:0.1, v/v/v) by consecutive dilution of
these solutions. A similar procedure has been performed for the
preparation of -blockers calibration standards from WS.
2.2. Sampling
Environmental matrices samples were collected from Pierre
Benite STP (downstream from Lyon, France). This STP receives
mainly urban and hospital wastewaters. For environmental
analysis, 24 h average samples were collected, whereas direct
samplings were used as matrix for method development. Both
influent and effluent samples were filtered up to 0.45 m on
membrane filters provided by Millipore (Molsheim, France) and
stored at 4 C. Samples were extracted within a maximum of 3
days after the collect.
2.3. Sample preparation
Samples were extracted on Oasis MCX cartridges (60 mg,
3 cc) from Waters (Saint Quentin en Yvelines, France). The solid
phase extraction was performed on an Aspec XL4 equipped with
a 404 syringe pump and an 8 position valve mate from Gilson
(Villiers le Sec, France), controlled by the 735 sampler Gilson
software.
Each 400 mL environmental sample was acidified with
400 L of FOA (85%), then spiked by the volume needed
(<50 L) of calibration standard for the standard addition
method. The sample was then loaded at a flow rate of
15 mL min1 using Aspec XL4 onto an Oasis MCX extraction
cartridge preconditioned with 3 mL of MeOH and 3 mL of water
acidified with 0.1% of FOA (85%). The cartridge was then rinsed
with 1.5 mL of MeOH/water/FOA (10:90:0.1, v/v/v) mixture.
Corticosteroids were eluted with 1 mL of MeOH/water/FOA
(70:30:0.1, v/v/v). A second wash with 1 mL of MeOH/FOA
(100:0.1, v/v) was performed. Finally, -blockers were eluted
with MeOH/ammonia solution (95:5, v/v). Consequently, 2
extracts were collected separately: the first one containing corticosteroids and the second one containing -blockers. Each wash
and elution step were performed at a flow rate of 0.3 mL min1 .

Table 1
Structures and mass spectrometric parameters used for the analysis of corticosteroids
Product ion
(m/z)

Betamethasone (Be) [378-44-9]

393.0

373.1

Budesonide (Bu) [51333-22-3]

431.0

Hydrocortisone (Hy) [50-23-7]

Structure

Orifice voltage
(V)

Ring voltage
(V)

Collision energy
(eV)

Retention time
(min)

41

240

13

6.3

413.3

140

15

11.0

362.9

327.2

31

200

29

3.8

Cortisone (C) [53-06-5]

361.0

163.1

26

160

31

4.1

Dexamethasone (D) [50-02-2]

393.0

373.1

41

240

13

6.5

A. Piram et al. / Talanta 74 (2008) 14631475

[M + H]+
(m/z)

Molecule [CAS]

1465

1466

435.0

321.1

150

19

7.9

Fluocinolone acetonide (Fc) [67-73-2]

453.1

413.2

11

160

17

8.7

Prednisolone (Pl) [50-24-8]

360.9

343.1

11

140

15

3.6

Prednisone (P) [50-03-2]

358.9

313.0

21

180

17

3.8

Triamcinolone (T) [124-94-7]

391.2

375.0

46

270

15

2.4

Triamcinolone acetonide (TA) [76-25-5]

435.0

415.1

150

15

7.6

A. Piram et al. / Talanta 74 (2008) 14631475

Flunisolide (Fs) [3385-03-3]

A. Piram et al. / Talanta 74 (2008) 14631475

1467

Table 2
Structures and mass spectrometric parameters used for the analysis of -blockers
[M + H]+
(m/z)

Product
ion (m/z)

Orifice
voltage (V)

Ring
voltage (V)

Collision
energy (eV)

Retention
time (min)

Acebutolol (Ac) [37517-30-9]

337.0

116.2

21

150

29

7.9

Alprenolol (Al) [13655-52-2]

249.9

116.2

16

140

23

9.2

Atenolol (At) [29122-68-7]

267.0

145.1

21

150

33

3.8

Bisoprolol (Bi) [66722-44-9]

326.1

116.2

21

140

25

8.7

Metoprolol (M) [51384-51-1]

268.0

116.3

21

150

23

8.1

Nadolol (N) [42200-33-9]

310.1

254.1

21

140

23

6.6

Pindolol (Pi) [13523-86-9]

248.9

116.2

16

130

23

7.5

Propranolol (Pr) [525-66-6]

259.9

116.3

16

140

23

9.1

Sotalol (S) [3930-20-9]

272.9

255.1

120

17

3.9

Timolol (Ti) [26-839-75-8]

317.0

261.1

16

140

23

7.8

Molecule [CAS]

Structure

1468

A. Piram et al. / Talanta 74 (2008) 14631475

Both eluates were evaporated to dryness in a Speed Vac from

Savant (Ramsey, Minnesota, USA). The dry residues were dissolved in 400 L of a mixture AcN/water (25:75, v/v). An
aliquot of 100 L was then injected in the LC/MS/MS device.
2.4. Liquid chromatographytandem mass spectrometry
The HPLC device consisted of Agilent 1100 series pump and
autosampler (Massy, France). As corticosteroids and -blockers
possess very different physicochemical properties (regarding
their hydrophobicity or chemical functions), they could not
be analysed on a same chromatographic column. -Blockers
were chromatographied on a 5 m Hibar Purospher Star column
(250 mm 4.6 mm I.D.) from Merck (Lyon, France), operating at 1 mL min1 . A gradient elution was performed using
two solvents: A1 was formate buffer (10 mM ammonium formate acidified to pH 3.8 with FOA) and solvent B was AcN
(Table 3a). Corticosteroids were chromatographied on a 3.5 m
C18 X-Bridge column (150 mm 2.1 mm I.D.) from Waters,
operating at 0.3 mL min1 . A gradient elution was performed
using two solvents: A2 was ultra-pure water and solvent B
was AcN (Table 3b). Both chromatographic separations were
performed at room temperature.
The HPLC device was coupled to a Sciex API 300 triple
quadrupole mass spectrometer from MDS Sciex (Toronto,
Canada) equipped with a TurboIonspray Source (TIS) operating in positive ion mode. Instrument control, data analysis and
processing were performed using the associated Analyst 1.4.1
software. The mass spectrometer was initially calibrated using
polypropylene glycol as standard (Applied Biosystems, Foster
city, USA), setting the resolution, as peak width at half height,
in the range 0.7 0.1 amu.
The nebulizer (zero air) and the curtain gas flow (nitrogen)
were set at 10 arbitrary units. The TIS source was operating at
400 C for corticosteroids and at 500 C for -blockers, with the
auxiliary gas flow (zero air) set at 8 L min1 . The TIS voltage
was set at 5000 V. As both peak width and number of MRM transitions differed in these two analytical methods, dwell times were
determined in order to define chromatographic peaks with about
20 points. Then, the dwell time was 100 and 50 ms for corticosteroids and -blockers, respectively. Other mass spectrometric
parameters are summarised in Table 1 for corticosteroids and
Table 2 for -blockers. These parameters were optimised using
the autotune feature of the Analyst 1.4.1 software, by infus-

ing each molecule separately. Product ions used for monitoring

were selected based on their significance in the MS/MS spectra.
For this study only one MRM (Multiple Reaction Monitoring)
transition was used although a second transition is generally
required to confirm the presence of an analyte by using the ratio
MRM1/MRM2. As the mass spectrometer used in this study did
not permit to work with large number of MRM transitions, it
was not possible to analyse a large number of compounds with
two MRM transitions. So we decided to use only one transition. During method development we noticed that when a peak
was observed on STP sample extracts, at the required retention
time, with the first MRM transition, the presence of the analyte was always confirmed by the second MRM transition. This
methodology was considered as acceptable for method development. However, consecutive routine analysis will require an
organisation of MRM transitions by retention time windows, in
order to increase the number of followed MRM transitions up to
two transitions per analyte, as lauded by the European directive
96/23/EC.
3. Results and discussion
3.1. Liquid chromatographytandem mass spectrometry
analysis
Tandem mass spectrometry is the best detection tool for environmental analysis because of its high sensitivity and selectivity,
especially in the MRM mode. Moreover, operating the mass
spectrometer in the MRM mode allows a shorter chromatography race since a perfect separation of all the compounds is
not required using this acquisition mode. Nevertheless, some
important aspects have to be considered in multiresidue method
development and in some cases, chromatographic separation
may be still required: (1) compounds measured at the same
MRM transition (e.g. betamethasone and dexamethasone corticosteroids which are diastereroisomers and have the same
precursor/product ion transition); (2) compounds which produce similar (overlapping) product ions upon fragmentation;
(3) compounds that are measured using the same product ion
but with different precursor ions, in order to avoid the risk of
a cross talk phenomenon, which is the case of most -blockers
with a m/z 116 product ion (acebutolol, alprenolol, bisoprolol,
metoprolol, pindolol and propranolol) and some corticosteroids (prednisolone/cortisone and budesonide/fluocinolone

Table 3
HPLC solvent gradient for the separation of (a) -blockers and (b) corticosteroids
(a)

(b)

Total time (min)

Flow rate

0
2
5
10

1
1
1
1

a
b
c

A1 = formate buffer.
B = can.
A2 = ultra-pure water.

(mL min1 )

%A1a

%Bb

Total time (min)

Flow rate (mL min1 )

%A2c

%Bb

85
85
55
55

15
15
45
45

0
6
9
20

0.3
0.3
0.3
0.3

70
60
0
0

30
40
100
100

A. Piram et al. / Talanta 74 (2008) 14631475

1469

Fig. 1. Mass spectrometric responses of 2 representatives corticosteroids: betamethasone (Be) and cortisone (C), and 2 representatives -blockers: propranolol (Pr)
and sotalol (S).

acetonide have a similar product ion at m/z 343 and 413,

respectively).
The couples of molecules quoted above have to reach the
mass spectrometer separately, and to be previously separated by
chromatography.
Thus, it is important to develop a chromatographic separation capable of separating these critical pairs. Although an
enrichment factor of 1000 can be expected from the sample
preparation, each parameter susceptible to lower the quantification limits is important to consider, such as mobile phase
composition. Then, it is also interesting to improve the mass
spectrometric sensitivity by using chromatographic mobile
phases able to increase the analyte ionization in the mass spectrometer ion source. Indeed, the organic solvent or the added
salts are susceptible to interfere with the ionization process in
the source of the mass spectrometer and affect the analytical
method sensitivity. So the quantification limits were lowered
by studying the influence of the mobile phase composition on
the analytes mass spectrometric response and it is possible to
detect environmental concentration in the ng L1 range, as it
was reported in literature [528,3033].
Chromatographic mobile phases were then chosen in order
to facilitate the analyte ionization in the source of the mass
spectrometer. Each compound was analysed in Flow Injection
Analysis mode (FIA) by testing different combinations of hydroorganics mixtures. MeOH and AcN were tested with different
aqueous phases: ultra-pure water, acidified water with 0.1%
of FOA, formate buffer (10 mM ammonium formate acidified
with FOA to pH 3.8) and basified water (ammonia solution
addition up to pH 11). The analyte responses of 4 representative compounds of both families are presented in Fig. 1. (2
corticosteroids: betamethasone and cortisone, in white; and 2
-blockers: propranolol and sotalol, in grey.)
For corticosteroids the best responses were observed with
mixtures of unmodified water with MeOH or unmodified water
with AcN. Therefore, the chromatographic gradient separation
was performed with water/AcN mixtures (Table 3b). The result-

ing chromatogram is shown in Fig. 2. The critical analyte couples

were separated within less than 12 min, excepted betamethasone
and dexamethasone. Indeed it was not possible to achieve a satisfactory resolution for these two compounds since their chemical
structures are very narrow (diastereoisomers), which results in
similar interactions with the chromatographic stationary phase.
These molecules will subsequently be quantified together. The
chromatography of budesonide resulted in two chromatographic
peaks, due to the presence of two epimers.
In the case of -blockers, the use of MeOH as organic solvent gave better responses. However, this solvent did not fit
with chromatographic separation criteria. Indeed, MeOH generated important peak tailing whatever the chromatographic
column tested. Thus, AcN should be used as organic constituent.
Although unmodified water gave the best results, it has not been
chosen because of the risk of poor robustness of the chromatographic separation. Finally, the best compromise between a good
chromatographic separation and a satisfactory analyte response
was to use a mixture of formate buffer and AcN or acidified
water (0.1% FOA) and AcN. A formate buffer/AcN gradient
(Table 3a) was selected. This mobile phase composition allowed
a satisfactory separation in less than 10 min as shown in Fig. 3.
3.2. Solid phase extraction
Solid phase extraction efficiency (purification and enrichment factor) is linked to a large number of parameters such
as flow rates, composition, and elution solvent volumes used
in each step of the procedure. Studying each parameter separately results in a high number of experiments, which should be
time consuming. Therefore, evaluating simultaneously the influence of several parameters is interesting for method development
and optimisation. Here is the advantage of chemometrics, which
allow the evaluation of several parameters by a restricted number
of experiments. This approach first requires a small knowledge
about the retention of target molecules on the SPE support. This
information can be obtained from a preliminary study.

1470

A. Piram et al. / Talanta 74 (2008) 14631475

Fig. 2. LC/MS/MS chromatogram of a 5 ng injected corticosteroids solution (see abbreviations in Table 1).

Corticosteroids are non-polar molecules whereas -blockers

are basic molecules with pKa s around 9. According to these
properties, in order to have a single sample preparation, a polymeric mixed-mode cation exchange reverse phase sorbent was
selected: Oasis MCX. Indeed, with apolar interactions MCX
sorbent is able to retain apolar molecules (such as corticosteroids) and with ionic interactions they can retain cations (such
as -blockers when protonated).
The evolution of the extraction yield of target analyte with the
composition of the eluant mixture was evaluated. The volume of
elution solvent used was fixed to 1 mL. In acidified MeOH/water
mixtures, corticosteroids extraction yield varied with the polarity
of the molecules. It has been concluded that (1) the composition of the eluant mixture should be less than 20% MeOH in
order to rinse the cartridge without eluting triamcinolone, the
less retained corticosteroid and (2) the percentage of MeOH
used to elute corticosteroids has to be higher than 60% in order
to have a satisfactory elution of budesonide, the most retained
corticosteroid. Moreover, no elution of -blockers was observed
in acidified mixtures. These results were interesting as it allowed
a second wash of the cartridge with 100% of acidified MeOH.
Indeed, acidic pH condition maintains these molecules under
their acidic structures, thus, they keep fixed on ion exchange

sites, whereas in basic media, -blockers become neutral and are

only subjected to apolar interactions. To remove all -blockers
from the cartridge, 95% of basified MeOH was necessary as
pindolol was not eluted with lower MeOH percentages.
These preliminary experiments showed that corticosteroids
and -blockers could be extracted from environmental samples with a unique solid phase procedure. Indeed, from the
same sample loading on the cartridge, corticosteroids fraction
is firstly purified then eluted. And secondly, -blockers fraction
was washed then eluted on this same cartridge.
Previous results can be strongly improved. Indeed, it does not
take care of important parameters such as the influence of the
volume loaded into the cartridge, nor evaluate the purification
efficiency, which could be evaluated by measuring matrix effect.
An improvement of the method on environmental samples was
then required.
To improve the SPE protocol, all the experiments were performed on Pierre Benite STP influents in order to cover the worst
scenario since untreated STP waters are expected to be the dirtiest matrix found in the environment and consequently the most
problematic.
The nine following parameters were firstly studied: the flow
rate used to percolate the sample on the cartridge, the sample vol-

A. Piram et al. / Talanta 74 (2008) 14631475

1471

Fig. 3. LC/MS/MS chromatogram of a 100 pg -blocker solution (see abbreviations in Table 2).

ume loaded, the eluant volume and methanol percentage for the
corticosteroids wash step, the eluant volume and methanol percentage for the corticosteroids elution step, the acidified MeOH
volume used to wash the -blockers fraction, the eluant volume used to eluate -blockers and its percentage of ammonia.
This study, which would not be detailed in this paper, allowed
to fix 5 critical parameters. The flow rate for loading (from 5 to
15 mL min1 ) did not influence neither the retention of target
analytes nor interferences and was then fixed to 15 mL min1 in
order to shorter the sample preparation time. The sample volume percolated was studied from 100 to 1000 mL and did not
affect the extraction yield of target molecules. It was then fixed
to 400 mL (leading to a 400 L extract) in order to perform
until 4 chromatographic injections per sample. The increase
of the volume percolated to clean up the -blockers fraction
(15 mL) did not increase the purification, and subsequently it
was maintained to 1 mL of MeOH acidified with 0.1% FOA.
This wash permits to eluate the remaining apolar interferents
from the cartridge, allowing to achieve a high purification status
for -blockers extracts. The best extraction yields for -blockers
were obtained by eluting with 1.5 mL MeOH/ammonia solution
28% (95:5, v/v), when 0.51.5 mL of MeOH/ammonia solutions
(98:295:5, v/v) were tested. These conditions gave satisfactory
purifications.

The 4 remaining parameters were studied more precisely,

thanks to a screening design (as shown in Table 4). The eluant volume for the corticosteroids wash step (called X1 ) varied
between 0.5 mL (level 1) and 1.5 mL (level +1); its composition (called X2 ) was 10% MeOH (level 1) or 15% MeOH (level
+1). The eluant mixture volume used to elute corticosteroids
(called X3 ) varied between 0.5 mL (level 1) and 1.5 mL (level
+1); its composition (called X4 ) was 65% MeOH (level 1) or
70% MeOH (level +1).
Table 4
22 factorial design of the experiment
Exp.

Step
1st wash

1
2
3
4
5
6
7
8

1st elution

V (mL)

%MeOH

V (mL)

%MeOH

0.5
1.5
0.5
1.5
0.5
1.5
0.5
1.5

10
10
15
15
10
10
15
15

0.5
1.5
1.5
0.5
1.5
0.5
0.5
1.5

65
65
65
65
70
70
70
70

1472

A. Piram et al. / Talanta 74 (2008) 14631475

Table 5
Corticosteroids extraction yields measured for the design of experiments (%)
Exp.

Pl

Fc

Be-D

Hy

Bu

Fs

TA

1
2
3
4
5
6
7
8

6
82
16
1
79
74
7
61

5
70
13
1
75
69
5
44

5
70
8
1
91
86
12
72

3
66
15
1
81
73
6
75

13
84
23
7
79
84
10
65

7
74
16
1
86
77
6
61

8
82
19
1
92
73
8
104

2
38
10
6
129
52
18
19

8
52
10
3
81
90
7
48

6
59
14
4
92
90
5
57

For each analyte 2 responses were considered. Firstly the

extraction yield was calculated following equation:
Yield (%) =

AreaA
100
AreaB

(1)

where AreaA is the analyte chromatographic area measured

when a spiked matrix was extracted and AreaB is the analyte
chromatographic area measured when an unspiked matrix was
extracted then its eluate was spiked.
Secondly, the matrix effect was measured following equation:


AreaB
Matrix effect (%) =
1 100
(2)
AreaC
where AreaC represents the analyte chromatographic area in
pure solvents.
Certainly the presence of matrix components in the extract,
when compared to standard solution in pure solvent, can cause
many errors, which can lead to inaccurate results. These matrix
components can promote either ion suppression or improvement
of the analyte ionization in the electrospray interface [34], and
should be removed whenever possible.
The extraction yield was headlined and matrix effect was secondarily considered. The extraction yields obtained for each run
of experimental design are shown in Table 5. Part of the experi-

ments has been duplicated. The experiments were reproducible,

except those which gave very poor extraction yields (under 5%)
but this cannot misshape the interpretation. In Fig. 4, the effects
of the screening model (yield = b1 X1 + b2 X2 + b3 X3 + b4 X4 ) are
presented. The higher the absolute value of the effect (b1 , b2 ,
b3 , b4 ), the higher the influence of the corresponding parameter on the response. As showed in Fig. 4, each tested parameter
has a strong influence on the extraction yield, since the output effects values are meaningful. The signal of the measured
response ( or +) is directly linked with the sign ( or +) of the
experimental design parameter tested. b1 , b3 and b4 are positive
for most analyte, thus the levels +1 of the parameters X1 , X3
and X4 increase the response. b2 is negative, thus the parameter X2 should be fixed to the level 1 in order to increase the
response. In other words, washing with 1.5 mL (X1 = +1) with an
acidified 10% MeOH mixture (X2 = 1) and eluting with 1.5 mL
(X3 = +1) of an acidified 70% MeOH mixture (X4 = +1) resulted
in increased extraction yields.
In order to decrease the eluant mixture volume, a final
improvement was performed by using the optimisation module
of Modde v.7 software (from Umetrics, Umea, Sweden). This
study shows that it is possible to obtain a satisfactory elution of
all corticosteroids with X3 = 0, i.e. by eluting with only 1 mL
of MeOH/water/FOA (70:30:0.1, v/v/v).

Fig. 4. Influence of the tested parameter on the extraction yield. b1 : volume used for the first wash; b2 : % MeOH used for the first wash; b3 : volume used for the first
elution; b4 : % MeOH used for the first elution.

A. Piram et al. / Talanta 74 (2008) 14631475

Table 6
Linearity range and IDL for the LC/MS/MS analysis
Molecule
Corticosteroids
Betamethasone + dexamethasone
Budesonide
Cortisol
Cortisone
Flunisolide
Fluocinolone acetonide
Prednisolone
Prednisone
Triamcinolone
Triamcinolone acetonide
-Blockers
Acebutolol
Alprenolol
Atenolol
Bisoprolol
Metoprolol
Nadolol
Pindolol
Propranolol
Sotalol
Timolol

Linearity range
(g L1 )

1473

3.3. Method performance evaluation

Correlation
coefficient

IDL
(pg)

1400
1100
4400
1400
1100
1100
1400
1400
1100
1100

0.999
0.999
0.993
0.998
0.999
0.999
0.992
1.000
0.996
0.998

32.0
0.7
7.7
1.3
2.5
0.6
9.1
6.7
77.3
0.5

0.1400
0.1100
0.1400
0.1100
0.1400
0.1400
0.1400
0.1100
0.1400
0.1400

0.992
0.995
0.993
0.985
0.990
0.991
0.992
0.990
0.991
0.989

0.3
1.0
0.4
0.1
0.5
0.3
0.3
0.2
2.0
0.9

The final protocol was described in detail in Section 2.

Moreover, we noticed that each experiment resulted in important matrix effects, except those that gave extraction yields
under 10%. To avoid matrix effect higher purification would
have to be reached; unfortunately we have shown that it could
not be done without strongly decreasing the extraction yield.
Another method consists in compensate matrix effect by working with deuterated compounds for each analyte. In this case
the isotopic molecule undergoes the same matrix effect than
the natural compounds, thus the chromatographic area ratio
analyte/deuterated compound results unchanged and is independent of matrix effect. However, this method is very expensive,
and moreover most studied deuterated molecules are not commercially available. As matrix effect could not be completely
eliminated, it was necessary to quantify unknown samples by
the standard addition method.

Calibration plots were obtained by standard solution analysis at eight concentrations for each analyte. The linearity of the
method was evaluated between 1 and 400 g L1 . To determine
the best weighting factor, concentrations were back-calculated
and the residual plot examined. The model with the lowest
bias and the most constant variance across the concentration
range was considered the best suited. A linear regression (1/x2
weighted for -blockers and 1/x for corticosteroids) gave the best
fit for the concentration/detector response with R2 values above
0.99 and residues less than 15% for each analyte in the concentration range tested (Table 6). As no blank matrix was found
out within the tested samples, it was not possible to evaluate
the method sensibility on environmental samples. Instrumental detection limits (IDL) were thus calculated as the minimum
injected mass giving a signal to noise ratio 3. The IDLs ranged
0.1 and 77.3 pg analyte injected, as shown in Table 6.
Instrumental precision was assessed by replicate injection
(n = 5) of the calibration solution in the same day. The intraday variance, expressed as relative standard deviation (R.S.D.%)
for each analyte peak area was usually below 15%. The
accuracy of the method was determined by recovery studies
conducted on spiked environmental samples (STP influents) at
500 ng L1 . Extraction yields and their R.S.D. are presented in
Table 7.
The analytical protocol was not fully validated since the aim
of this work was to develop an extraction procedure and test
it on environmental samples, in order to investigate if target
molecules were found in the environment.
3.4. Environmental analysis
Influent samples from Pierre Benite STP have been extracted
and analysed as described above, by the standard addition
method. With this protocol, the environmental sample is divided
in 4 aliquots of 400 mL. Each aliquot was spiked with a relevant
quantity of standard (up to twice the suspected environmental
concentration). The linearity of standard addition was verified
and allowed a reliable environmental quantitation. The quantities measured are reported in Table 8. The studied -blockers
have been detected with concentrations up to 318 ng L1 for

Table 7
Extraction yields and relative standard deviation in this study, on STP influents (n = 5)
Corticosteroid

Yield (%)

R.S.D. (%)

-Blocker

Yield (%)

R.S.D. (%)

Betamethasone
Budesonide
Cortisone
Cortisol
Dexamethasone
Flunisolide
Fluocinolone acetonide
Prednisolone
Prednisone
Triamcinolone
Triamcinolone acetonide

67
86
71
70
75
54
84
62
55
63
70

10
12
10
22
11
19
24
16
13
11
10

Acebutolol
Alprenolol
Atenolol
Bisoprolol
Metoprolol
Nadolol
Pindolol
Propranolol
Sotalol
Timolol

68
61
68
60
62
60
10
66
68
62

4
6
24
7
7
9
24
7
9
9

1474

A. Piram et al. / Talanta 74 (2008) 14631475

Table 8
Measured concentration in Pierre Benite influents/effluents STP
Compound

Influents (ng L1 )

Effluents (ng L1 )

Corticosteroids
Betamethasone + dexamethasone
Budesonide
Cortisone
Cortisol
Flunisolide
Fluocinolone acetonide
Triamcinolone
Triamcinolone acetonide

15
n.d.
174
53
7
0.3
31
40

7
3
229
63
5
11
30
3

-Blockers
Acebutolol
Atenolol
Bisoprolol
Metoprolol
Nadolol
Pindolol
Propranolol
Sotalol
Timolol

66
170
73
111
10
2
92
318
32

0.4
1.8
1
1
0.1
0.4
2
65
58

sotalol. The measured -blockers concentrations are in accordance with the amounts reported in literature [13,2125]. All
the studied corticosteroids were detected, excepted budesonide.
Concentrations reach 174 ng L1 for cortisone. Cortisone and
cortisol which are not only prescription drugs but also endogenous steroids are continuously discharged in sewages. So it is
not surprising these 2 compounds were present to higher concentrations than other corticosteroids.
Then, the developed protocol was applied, to Pierre Benite STP effluent samples (results shown in Table 8). Our aim
was to have a first estimation on the quantities discharged by
STP effluents in environmental waters, i.e. to know if the treatment was able to totally remove corticosteroids from water. All
the target compounds have been detected, including budesonide
which had not been detected in influent samples. This can be
due to the hydrolysis of budesonide conjugates during the STP
treatment. This study demonstrates that corticosteroids reach
STPs and are continuously discharged by them in environmental
waters.
To our knowledge, this is the first study that demonstrates
contamination of environmental waters by corticosteroids.
4. Conclusions

concentrations in Europe or America [528]. It also permitted

the detection of corticosteroids in STP influents and effluents.
This demonstrates that these molecules are a matter of concern
in an environmental context and should be part of monitoring
studies.
As both corticosteroids and -blockers were detected on
tested environmental samples, further work will include a full
procedure validation and its application on a large number of
samples.
Acknowledgements
The authors acknowledge the Region Rhone-Alpes for its
financial support and Grand Lyon for assistance on sampling.
The authors thank also Sanofi-Aventis for its materials help.
References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]

A quantitative analytical procedure has been developed to

study 11 corticosteroids and 10 -blockers in environmental
waters. The use of mixed-mode MCX cartridges allows a single and specific sample preparation. As far as we know, this is
the first procedure proposed for the environmental quantification of corticosteroids. Furthermore, it gives a high purification
of -blockers, compared to existing methods. Indeed a 100%
MeOH wash is allowed for these compounds with no loss of
analytes.
This method was applied on environmental samples and the
-blockers concentrations were in accordance with reported

[22]
[23]
[24]
[25]
[26]
[27]
[28]

M. Cleuvers, Ecotoxicol. Environ. Saf. 59 (2004) 309.

M. Cleuvers, Toxicol. Lett. 142 (2003) 185.
M. Cleuvers, Chemosphere 59 (2005) 199.
K. Fent, A.A. Weston, D. Caminada, Aquat. Toxicol. 76 (2006)
122.
D. Ashton, M. Hilton, K.V. Thomas, Sci. Total Environ. 333 (2004)
167.
A.C. Berlfroid, A. Van der Host, A.D. Vehtaak, A.J. Schafer, G.B.J. Rijs,
J. Wegener, W.P. Cofino, Sci. Total Environ. 225 (1999) 101.
J.D. Cahill, E.T. Furlong, M.R. Burkhardt, D. Kolpin, L.G. Anderson, J.
Chromatogr. A 1041 (2004) 171.
J.G.M. Derksen, G.B.J. Rijs, R.H. Jongbloed, Water Sci. Technol. 49 (2004)
213.
T. Heberer, J. Hydrol. 166 (2002) 175.
T. Heberer, Toxicol. Lett. 131 (2002) 5.
K.V. Thomas, M.J. Hilton, Mar. Pollut. Bull. 49 (2004) 213.
E. Zuccato, S. Castiglioni, R. Fanelli, J. Hazard. Mater. 122 (2005)
205.
M.J. Hilton, K.V. Thomas, J. Chromatogr. A 1015 (2003) 129.
O.A.H. Jones, N. Voulvoulis, J.N. Lester, Water Res. 36 (2002) 5013.
S. Castiglioni, R. Bagnati, D. Calamari, R. Fanelli, E. Zuccato, J. Chromatogr. A 1092 (2005) 206.
D. Bendz, N.A. Paxeus, T.R. Ginn, F.J. Loge, J. Hazard. Mater. 122 (2005)
195.
P.E. Stackelberg, E.T. Furlong, M.T. Meyer, S.D. Zaugg, A.K. Hendersen,
D.B. Reissmen, Sci. Total Environ. 329 (2004) 99.
D.W. Kolpin, E.T. Furlong, M.T. Meyer, S.D. Zaugg, L.B. Barber, H.T.
Buxton, Environ. Sci. Technol. 36 (2002) 1202.
M.J. Lopez de Alda, D. Barcelo, J. Chromatogr. A 892 (2000) 391.
A. Salvador, C. Moretton, A. Piram, R. Faure, J. Chromatogr. A 1145 (2007)
102109.
D.B. Huggett, I.A. Khan, C.M. Foran, D. Schlenk, Environ. Pollut. 121
(2003) 199.
R. Andreozzi, M. Raffaele, P. Nicklas, Chemosphere 50 (2003) 1319.
T.A. Ternes, Water Res. 32 (1998) 3245.
D. Calamari, E. Zuccato, S. Castiglioni, R. Bagnati, R. Fanelli, Environ.
Sci. Technol. 37 (2003) 1241.
C. Mi`ege, M. Favier, C. Brosse, J.P. Canler, M. Coquery, Talanta 70 (2006)
739.
M. Petrovic, E. Eljarrat, M.J. Lopez de Alda, D. Barcelo, J. Chromatogr.
A 974 (2002) 23.
E.P. Kolodziej, T. Hareter, D.L. Sedlak, Environ. Sci. Technol. 38 (2004)
6377.
K.V. Thomas, M.R. Hurst, P. Matthiessen, M. McHugh, A. Smith, M.J.
Waldock, Environ. Toxicol. Chem. 21 (2002) 1456.

A. Piram et al. / Talanta 74 (2008) 14631475

[29] H. Sanderson, D.J. Johnson, T. Reitsma, R.A. Brain, C.J. Wilson, K.R.
Solomon, Regul. Toxicol. Pharmacol. 39 (2004) 158.
[30] T.A. Ternes, TrAC 20 (2001) 419.
[31] D. Bendz, N.A. Paxeus, T.R. Ginn, F.J. Loge, J. Hazard. Mater. 122 (2005)
195.

1475

[32] H. Budzinski, A. Togola, Spectra Analyse 33 (2004) 1.

[33] L. Amendola, F. Molaioni, F. Botr`e, J. Pharm. Biomed. Anal. 23 (2000)
211.
[34] W.A. Kormacher (Ed.), Using Mass Spectrometry for Drug Metabolism
Studies, CRC press, Boca Raton, FL, 2005, p. 103.