Environment International 35 (2009) 997–1003

Contents lists available at ScienceDirect

Environment International
j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / e n v i n t

Analysis of phytoestrogens, progestogens and estrogens in environmental waters

from Rio de Janeiro (Brazil)
M. Kuster a,⁎, D.A. Azevedo b, M.J. López de Alda a, F.R. Aquino Neto b, D. Barceló a,c
a
Institut de Diagnòstic Ambiental i Estudis de l'Aigua (IDAEA), Spanish Council for Scientific Research (CSIC), Barcelona, Spain
b
Departamento de Química Orgânica, Instituto de Química, Universidade Federal do Rio de Janeiro, Rio de Janeiro, Brazil
c
Catalan Institute for Water Research (ICRA), Parc Científic i Tecnològic de la Universitat de Girona, Girona, Spain

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

Article history: The environment is currently exposed to a large variety of man-made chemicals (e.g. for industrial, medicinal
Received 11 March 2009 use) which have potential adverse effects to its ecological status. In addition, the densely populated areas
Accepted 21 April 2009 represent local high emissions of those chemicals leading to more aggravating consequences. Estrogenic
Available online 20 May 2009
compounds that end-up in environmental water directly affect living organisms by interfering with their
endocrine metabolism. The assessment of their presence in the environment requires sensitive and selective
Keywords:
Phytoestrogens
analytical methods. Nineteen estrogenic compounds belonging to different classes (5 free estrogens, 6
Estrogens conjugated estrogens, 3 progestogens and 5 phytoestrogens) have been studied. The analytical methodology
Water analysis developed is based on solid phase extraction followed by liquid chromatography tandem mass spectrometry
LC–MS/MS and has been applied to study the occurrence of the above mentioned analytes in environmental waters from
Brasil the state of Rio de Janeiro (Brazil). Due to insufficient infra-structure in this region, waste waters are released
Rio de Janeiro onto the environment without or with incomplete previous treatment. The results show that high levels of
the phytoestrogens daidzein, coumestrol and genistein of up to 366 ng/L and progesterone of up to 47 ng/L
could be found in river water. Estrogens and their conjugated derivatives were detected in the lower ng/L
range up to 7 ng/L. The main estrogens estrone, estradiol and the synthetic ethinyl estradiol could not be
detected. The developed method showed overall good performance with recoveries above 80% (with one
exception), limits of detection ≤ 2 ng/L, good linearity and reproducibility.
© 2009 Elsevier Ltd. All rights reserved.

1. Introduction In order to gain more knowledge of the contaminants fate and

Over the past 10 years, many synthetic compounds as well as
natural (animal, human and plant) substances present in environ-
mental waters have been found to affect hormonal functions of
aquatic organisms (Halling-Sørensen et al., 1998). Among those
compounds the steroidal sex hormones (estrogens and progestogens),
used in large amounts for medicinal purpose, exhibit the highest
estrogenic activities, being strong endocrine disrupters. The nonster-
oidal polycyclic phenol phytoestrogens are plant compounds present
at high concentrations in soy, clover, alfalfa and other legumes. This
class of compounds can induce estrogenic or, in some cases, anti-
estrogenic effects (Hwang et al., 2006) and has found application in
the modern medicine as well. The main sources of input of these
compounds into environmental waters are the incomplete removal
during waste water treatment (Bacaloni et al., 2005; Kuster et al.,
2007), run-off from manure treated soils (Erbs et al., 2007; Hanselman
et al., 2003), and in the case of the phytoestrogens direct release
from the cultivated plant might also be possible (Erbs et al., 2007).
⁎ Corresponding author.
E-mail address: makqam@cid.csic.es (M. Kuster).
their potential risk to the studied environment it is essential to
determine therein their occurrence. Analytical methodologies with
high sensitivity and selectivity have to be employed. Several authors
have reported the analysis of the above mentioned compounds in
environmental waters. Currently the use of solid phase extraction
(SPE) followed by liquid chromatography (LC) and mass spectrometry
is consolidated for this purpose (Kuster et al., 2007; Lopez de Alda
et al., 2003). When utilizing tandem mass spectrometry (MS/MS)
higher confidence and lower detection limits can be achieved
(Bacaloni et al., 2005; Laganà et al., 2004; Reemtsma, 2001).
The concentrations found in natural waters in several European
countries are usually below 5 ng/L (Baronti et al., 2000; Belfroid et al.,
1999; Kuch and Ballschmiter, 2001; Kuster et al., 2008) for the
estrogens and progestogens. Considering that at 1 ng/L can cause
adverse effects (Desbrow et al., 1998; Hansen et al., 1998; Petrovic
et al., 2002; Routledge et al., 1998), such low levels are not to be
neglected due to the fact that little is known about synergistic and
long term effects. Estrone has been detected occasionally at levels
close to 20 ng/L (Kawaguchi et al., 2004; Xiao et al., 2001). Then again,
alarming high levels of up to 73 ng/L of estrogens have been found in
China/Taiwan (Chen et al., 2007).

0160-4120/$ – see front matter © 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.envint.2009.04.006
998 M. Kuster et al. / Environment International 35 (2009) 997–1003
Fig. 1. Map of the state of Rio de Janeiro with the sampling locations. 1, Resende; 2, Volta
Redonda; 3, Barra Mansa; 4, Seropédica; 5, Campos dos Goytacazes; 6, Duque de Caxias;
7, Rio de Janeiro city; 8, Lumiar.
Phytoestrogens, however, exhibit a lower estrogenic potency by up to
two orders of magnitude than the reference estradiol (Bovee et al., 2004).
Like other contaminants with similar estrogenic potency adverse effects
can be expected to occur in the environment and have already been
reported (Clotfelter and Rodriguez, 2006). The levels of phytoestrogens
reported are mostly below 10 ng/L (Bacaloni et al., 2005; Kang et al.,
2006; Laganà et al., 2004), similar to estrogens; however, Kawanishi et al.
(2004) found extremely high concentrations of 43 and 143 µg/L of
daidzein and genistein respectively in river water in Japan.
The objective of this work was to determine the levels of selected
estrogens, progestogens and phytoestrogens in waters from highly
populated areas in the state of Rio de Janeiro (Brasil) in order to have a
first overview of this type of contamination as, to the authors knowl-
edge, there is no previous report on the levels of such contaminants in
the studied area. The analytical method developed, based on solid phase
extraction (SPE) followed by liquid chromatographic separation and
tandem mass spectrometric detection (LC–MS/MS), allows the simul-
taneous analysis of a wide variety of important estrogenic compounds
from synthetic, animal and plant origin in environmental waters. Nine-
teen compounds were monitored with this method including 3 natural
and 2 synthetic free estrogens (estradiol, estrone, estriol, ethinyl
estradiol and diethylstilbestrol), 6 conjugated estrogens (sulfate and
glucuronide derivatives of estradiol, estrone and estriol), 1 natural and 2
synthetic progestogens (progesterone, norethindrone and levonorges-
trel) and finally 5 phytoestrogens (daidzein, resveratrol, coumestrol,
genistein and biochanin A). This methodology has been applied to the
analysis of 20 water samples, collected in two campaigns performed in
March and April 2007, from the Paraíba do Sul, Guandú, and Macaé
Rivers, from the Pavuna Channel and from two lagoons situated in the
south part of the city of Rio de Janeiro (see Fig. 1).
2. Materials and methods
2.1. Standards and reagents
Pure standards of the target estrogens, progestogens, phytoestro-
gens were purchased from Sigma-Aldrich (St Louis, MO, USA).
Individual stock solutions of the analytes were initially prepared at
1 or 2 mg/mL in methanol, except for coumestrol that was dissolved in
DMSO. Standard mixtures of the various compound classes were then
prepared in methanol at different concentrations by appropriate
dilution of the individual stock solutions.
For sample processing chromatographic grade methanol from
TediaBrazil (Rio de Janeiro, Brazil), ultra-pure water (Milli-Q)
obtained from a Millipore Direct-Q5 system (Bedford, MA, USA), and
hydrochloric acid from Merck (Darmstadt, Germany) were used.
LiChrolut RP-18 and Oasis HLB SPE cartridges were from Merck
(Darmstadt, Germany) and Waters (Milford, MA, USA) respectively.
For elution and further analysis HPLC-grade water, acetonitrile and
methanol, and hydrochloric acid, formic acid and ammonium
hydroxide were purchased from Merck (Darmstadt, Germany).
Nitrogen for drying (99.995% purity) was from Air Liquide (Spain).
2.2. Analytical methods
Water samples were collected in previously cleaned amber glass
bottles and transported under cooled conditions at 4 °C. Preceding
extraction samples were filtered through 0.45 µm Nylon filters
(Whatman, Maidstone, UK) and adjusted to pH 5 using hydrochloric
acid. The analytical procedure is schematically shown in Fig. 2.
Extraction was performed using a solid-phase extraction (SPE)
manifold from Waters (Milford, MA, USA). The target analytes were
extracted with LiChrolut RP-18 500 mg cartridges previously condi-
tioned with water and methanol. After loading of the 500 mL sample
the cartridge was washed with water (5 mL), dried for 15 min with
vacuum and stored at − 20 °C. Further processing was performed by
passing a small amount of hexane (500 uL) before elution of the
analytes. Subsequent elution of the analytes was performed in two
steps with 5 mL methanol and 5 mL methanol containing 2% ammo-
nium hydroxide. After drying under a N2-stream and reconstitution
with 250 µL methanol the extracts were analysed by LC–MS/MS.
LC–MS/MS analyses were carried out in a system consisting of a
Waters Alliance 2690 LC pump equipped with an auto-sampler with
the volume injection set to 20 µL (Waters, Milford, MA, USA) and
connected in series with the MS/MS detector.
Chromatographic separation was performed using a reversed-
phase Purospher STAR-RP-18e analytical column (125 × 2 mm, 5 µm
particle diameter) preceded by a guard column (4 × 4 mm, 5 µm) of
the same packing material from Merck (Darmstadt, Germany). The
chromatographic separation was performed at controlled room
temperature (20–22 °C) at a flow rate of 0.2 mL/min with a 45 min
gradient starting from 10% acetonitrile in water, increasing to 50%
acetonitrile in 5 min and continuing to 80% in 20 min. During the
following 5 min the column was cleaned with 100% acetonitrile,
readjusted to the initial conditions in 2 min, and equilibrated for
further 13 min. For phytoestrogens and progestogens chromatography
was performed with 0.1% formic acid in water.
Fig. 2. Complete scheme of the analytical method.
 M. Kuster et al. / Environment International 35 (2009) 997–1003 999

Table 1
Time scheduled SRM conditions used for the analysis of estrogens, progestogens and phytoestrogens.

Analyte Time window Retention time (min) SRM transitions (m/z) prec. ion→prod. ion Cone (V) Collision (eV)
Estrogens (NI)
Estriol-3-sulfate 0.00–11.00 7.78 367→287; 367→80 50 35; 30
Estriol-16-glucuronide 7.84 463→85; 463→287 35 30; 30
Estradiol-17-glucuronide 8.39 447→113; 447→271 40 30; 50
Estrone-3-glucuronide 8.60 445→113; 445→269 30 25; 30
Estradiol-3-sulfate 9.55 351→271; 351→80 40 35; 30
Estriol 9.87 287→171; 287→145 50 40; 40
Estrone-3-sulfate 10.24 349→269; 349→145 40 40; 40
Estradiol 11.00–25.00 13.04 271→145; 271→183 50 45; 45
Ethinyl estradiol 14.36 295→145; 295→159 50 40; 40
Estrone 14.50 269→145; 269→143 50 40; 45
Diethylstilbestrol 15.18 267→222; 267→237 30 35; 45

Progestogens and phytoestrogens (PI)

Daidzein 0.00–13.00 10.47 255→199; 255→137 45 25; 25
Resveratrol 10.95 229→135; 229→107 25 15; 15
Coumestrol 11.64 269→197; 269→213 50 25; 25
Genistein 11.65 271→91; 271→153 45 40; 25
Norethindrone 13.00–17.70 14.15 299→171; 299→145 30 20; 20
Biochanin A 15.42 285→213; 285→242 45 35; 30
Levonorgestrel 16.35 313→245; 313→185 30 20; 20
Progesterone 17.70–25.00 20.60 315→109; 315→97 30 25; 25

Detection was performed using a Quattro triple–quadrupole mass
spectrometer from Micromass (Manchester, UK), equipped with an
orthogonal electrospray (ESI) ionization source. Two different selected
reaction monitoring (SRM) transitions were monitored per compound:
the first and more abundant transition was used for quantitation and the
second transition for confirmation of the results (see Table 1). To
maximize sensitivity, data acquisition was performed under time-
scheduled conditions (time windows are shown in Table 1).
Other MS/MS experimental conditions were as follows: capillary
voltage, 3.5 kV; source temperature, 150 °C; desolvation temperature,
450 °C; extractor voltage, 2 V; and RF lens, 0.4 V. Nitrogen was used as both
the nebulising and the desolvation gas. The flow-rate of the nebulising gas
was set at 60 L h− 1, and that of the desolvation gas at 550 L h− 1. Argon was
used as collision gas with a pressure of 2.58×10− 3 mbar.
Instrument control, data acquisition and evaluation were per-
formed by Masslynx 4.0 software (Micromass, Manchester, UK).
2.3. Description of the sampling sites
The state of Rio de Janeiro is located in the south east of Brazil and
has approximately 15.4 million inhabitants, about 6 million inhabi-
tants live in the metropole of Rio de Janeiro. The main water sources
are the Paraíba do Sul and Gaundú Rivers, the latter receives large
amounts of water from Paraíba do Sul via transposition/permutation
in order to be able to supply the equivalent of 80% of the metropoles
water needs. A map showing the sampling sites position is shown in
Fig. 1. The sample codes and GPS coordinates are listed in Table 2.
For this work the choice of sampling sites was based on the
demographic density, economic activities and interest for the water
supply. In the south east of the state, sampling sites 1, 2 and 3 are
located in middle sized cities (Resende 128 000 inh., Volta Redonda
260 000 inh., Barra Mansa 176 000 inh.) along the Paraíba do Sul River
with metal and mechanical industries of great economical impor-
tance. Sampling site 4 is located in Seropédica where water from
Guandú River is abstracted for water treatment and further supplied
to the Rio de Janeiro metropole. At this site, river water as well as
treated water have been collected. In the north, at the estuary of the
Paraíba do Sul River in Campos dos Goytacazes city (430 000 inh.) is
located sampling site 5. In this region the main economical activities
are based on agriculture, the agro-industry, livestock breeding and
biotechnology.
The following sites are located in the metropolitan area of Rio de
Janeiro. Site 6 is located in the municipality of Duque de Caxias;
samples have been taken from Pavuna Channel. Several households in
this region do not have adequate sanitary infrastructure, as a
consequence, untreated waste water ends up in the many channels
that exist in the surroundings. In the city of Rio de Janeiro, samples
from two lagoons located in the south, lagoon of Jacarépaguá and of
Rodrigo de Freitas, have been collected. The metropolitan region is
rich in chemical, pharmaceutical and biotechnology industries. Finally,
a reference sample, a non-polluted sample was collected at site 8 from
the Macaé River in Lumiar, Nova Friburgo city, located in a moun-
tainous region.
Two sampling campaigns have been performed in the middle of
March and April of 2007. A total of 20 samples have been collected
(see Fig. 1).

Table 2
Sampling site details and GPS coordinates. 3. Results and discussion

Sample Location Source GPS coordinates (S/W) 3.1. Method development

1-R Resende Paraíba do Sul River 22″27.94′/44″26.90′
2-R Volta Redonda Paraíba do Sul River 22″30.05′/44″05.42′ The aim was to perform a single extraction procedure for all three compound
3-R Barra Mansa Paraíba do Sul River 22″31.50′/44″11.33′ groups. The influence on the recovery of each compound by variation of the cartridge
4-R Seropédica Guandú River 22″49.28′/43″37.31′ solid phase properties, the sample loading flow rate, the pH of the sample, the solution
4-T Seropédica Treated water 22″49.76′/43″36.99′ for the wash and the elution of the cartridge were evaluated. The optimization of the
5-R Campos de Goytacazes Paraíba do Sul River 21″44.47′/41″20.16′ method was performed using Milli-Q or HPLC water and a mixture of river and waste
6-R Duque de Caxias Pavuna Channel 22″48.05′/43″18.33′ water treatment plant (WWTP) effluent.
7-1-L Rio de Janeiro Rodrigo de Freitas Lagoon 22″57.79′/43″12.60′ Two of the most commonly used SPE cartridges were tested, a reversed phase C18
7-2-L Rio de Janeiro Jacarepaguá Lagoon 22″58.50′/43″23.08′ based cartridge (LiChrolut RP-18) and a co-polymeric cartridge (Oasis HLB). Similar
8-R Lumiar Macaé River 22″23.20′/42″18.54′ recoveries were obtained for both cartridges for all compounds, with exception of the
phytoestrogens resveratrol, daidzein, coumestrol and genistein, which showed an
1000 M. Kuster et al. / Environment International 35 (2009) 997–1003

Fig. 3. Recoveries at different sample pH.

increase of recovery of 15 to 50% using LiChrolut RP-18e. Hence the extraction method is improvements could be observed. Therefore, elution was performed in two steps, first
based on the use of LiChrolut RP-18e cartridges. The use of Oasis HLB is wide spread for with 5 mL MeOH and secondly with 5 mL MeOH containing 2% ammonium hydroxide.
the extraction of those compounds, however our results show better performance of The extracts were further analysed by LC-ESI-MS/MS. Electrospray ionization was
the reversed phase C18 cartridges. performed in the negative mode (NI) for the estrogens and in the positive mode (PI) for
A too fast flow rate during loading of the samples onto the SPE cartridge can result the progestogens and phytoestrogens. Therefore two chromatographic runs for each
in low recoveries due to breakthrough during the analyte retention. Most studies work ionization mode are carried out. Mass acquisition was performed in the selected
below the 10 mL/min flow margin (Erbs et al., 2007; Kang et al., 2006; Koh et al., 2007; reaction monitoring (SRM) mode by acquiring two SRM transitions per compound
Salste et al., 2007). Tests performed with 500 mL sample volume at 5, 10 and 15 mL/min (quantitation and confirmation). By means of on-column injections, the optimization of
showed no influence on the recovery of the analytes. Given that labour time is an cone and collision voltages could be done for each SRM transition (see Table 1). The
important aspect in environmental analysis loading the sample at 15 mL/min brings detection has been split in time windows in order to increase the response by keeping
several advantages and has been applied in this method. the total scan time of the quadrupole below 1 s (i.e. the sum of the dwell time of all SRM
The pH of environmental water samples may vary considerably and impair the transitions in one time window did not surpass 1 s). See Table 1 for the listing of the
extraction efficiency. A test to assess the extent of this effect was performed using monitored SRM transitions and the detailed parameters employed in MS acquisition.
spiked HPLC water (100 ng/L of target analytes) at pH 3, 5, 7, 9. Unlike the Oasis HLB In previous works studying estrogens and progestogens liquid chromatography was
cartridge, we observed lower extraction efficiency for the estrogens E2G, E1G, E3S, and carried out using acetonitrile and water as mobile phase and a gradient starting at 10%
EE2 at pH 7 using LiChrolut RP-18 cartridges (see Fig. 3). As a result the sample had to be ACN increasing to 80% in 20 min. However, poor ionization of the phytoestrogens was
adjusted to pH 5 prior to solid phase extraction.
An important step in the extraction procedure is the washing of the cartridge
following the loading of the compounds. Commonly water is used for this purpose,
Table 3
however a more effective and selective wash can improve the detection of the
Performance parameters.
compounds by eliminating interfering compounds from the extract. The use of
methanol water solutions (also acidified or with ammonium hydroxide) and small Analyte r2 RSD Rec HPLC Rec River LOD HPLC LOD River LOD increase
amounts of pure organic solvents (e.g. hexane, methanol) has been reported previously (%) (%) (%) (ng/L) (ng/L) (%)
(Bacaloni et al., 2005; Baronti et al., 2000; Chen et al., 2007; Gentili et al., 2002; Kang et
E3G 0.992 7 90 35 0.16 0.56 244
al., 2006; Reddy et al., 2005; Salste et al., 2007). The employment of the wash solutions
E2G 0.998 6 81 23 0.23 0.74 228
10% MeOH in water, 10% MeOH with 2% acetic acid, 10% MeOH with 2% ammonium
E1G 0.999 7 88 25 0.15 0.28 84
hydroxide, pure water, and 500 µL of hexane and MeOH was tested.
E3S 0.999 6 97 14 0.04 0.07 66
The use of 500 µL of MeOH resulted in the severe loss of most analytes (e.g., up to
E2S 0.999 7 153 19 0.04 0.15 282
50% loss of estrogens conjugates), whereas the use of 10% MeOH in water and 10%
E1S 0.998 4 87 28 0.36 0.16 − 56
MeOH with 2% ammonium hydroxide affected most phytoestrogens. The recoveries of
E3 0.992 6 89 19 0.41 1.13 172
genistein and biochanin A, for instance, were reduced by half when using 10% MeOH in
E1 0.994 6 92 59 0.57 1.15 103
water, and the average recoveries obtained for most phytoestrogens fell down to about
E2 0.997 5 82 38 1.22 2.27 86
20% when using 10% MeOH with 2% ammonium hydroxide. On the other hand 10%
EE2 0.993 6 87 47 1.51 7.55 400
MeOH containing 2% acetic acid slightly lowered the recoveries of estradiol and
DES 0.992 6 65 42 0.56 0.92 65
ethinylestradiol by about 20 and 10%, respectively. Finally the best option was to wash
DAI 0.998 10 136 94 0.49 0.95 89
the cartridge with 1 mL HPLC water followed by 500 µL of hexane.
RES 0.997 10 87 13 1.23 2.02 64
In the literature several solvents are reported to elute efficiently estrogens (MeOH/
COU 0.999 9 116 65 0.93 1.72 85
DCM (Baronti et al., 2000; Chen et al., 2007; Koh et al., 2007; Laganà et al., 2004), EtOAc
GEN 0.999 15 121 111 0.72 2.46 242
(Isobe et al., 2003; Reddy et al., 2005)). The elution of phytoestrogens has been
NOR 0.996 8 95 81 2.01 3.59 79
performed mostly using pure MeOH (Bacaloni et al., 2005; Erbs et al., 2007; Kang et al.,
BIO 0.998 11 122 138 1.01 1.73 71
2006). In previous works performed in our group the use of MeOH for eluting estrogens
LEV 0.988 10 102 101 0.80 1.16 45
and progestogens has been proven to be optimal (Rodriguez-Mozaz et al., 2004). For
PRO 0.999 9 106 108 0.41 0.39 −5
the current multianalyte procedure, including estrogens (free and conjugated),
progestogens and phytoestrogens, we investigated if the addition of 5 mM TEA and E3G, Estriol-16-glucuronide; E2G, Estradiol-17-glucuronide; E1G, Estrone-3-glucuronide;
2% ammonium hydroxide to MeOH improved the elution of the studied compounds. E3S, Estriol-3.sulfate; E2S, Estradiol-3-sulfate; E1S, Estrone-3-sulfate; E3, Estriol; E1,
Elution with methanol under basic conditions has been reported as advantageous by Estrone; E2, Estradiol; EE2, Ethinyl estradiol; DES, Diethylstilbestrol; DAI, Daidzein; RES,
several authors (Gentili et al., 2002; Isobe et al., 2003; Reddy et al., 2005). Ammonium Resveratrol; COU, Coumestrol; GEN, Genistein; NOR, Norethindrone; BIO, Biochanin A; LEV,
hydroxide increased the recovery of E3 and the conjugates. Otherwise no relevant Levonorgestrel; PRO, Progesterone.
 M. Kuster et al. / Environment International 35 (2009) 997–1003 1001

Table 4 The limits of detection (LODs) and quantitation (LOQs) of the method were
Summary of results. experimentally estimated from the analysis of spiked Milli-Q water and river samples as
the minimum concentration of analyte giving a signal-to-noise ratio of 3 and 8,
Compound Detected range Average concentration Positive samples respectively. As shown in Table 3, LODs in HPLC water were below 1 ng/L for 14 out of
(ng/L) (ng/L) (n = 20) 19 compounds. The lowest sensitivities were found for norethindrone, ethinyl estradiol,
E3G 0.34–2.67 1.97 7 resveratrol and estradiol (2.01, 1.51, 1.23 and 1.22 ng/L, respectively). The average of the
E2G 1.10–7.34 3.50 11 LODs calculated for each sample analysed in this work showed as expected increased
E2S 0.59–0.85 0.72 2 values (see Table 3). The most affected compounds are EE2, with a 400% increase, followed
E3 1.00–7.27 3.68 14 by E2S (282%), E3G (244%), GEN (242%), E2G (228%) and E3 (172%).
DAI 36.2–276 156 2 The method developed is also highly selective and with the monitoring of two SRM
COU 170 170 1 transitions per compound, which is essential to reduce the risk of false positive results
GEN 3.96–366 97.6 4 (Beck et al., 2005; Laganà et al., 2004), the minimum number of identification points
PROG 0.51–47.2 9.35 10 (3) required by the European Commission (2002/657/CE) for reliable identification
and quantification of organic residues in animals and fresh meat, a guideline that is
increasingly being adopted in the environmental field (Belfroid et al., 1999), is achieved.
For the positive confirmation of analytes in the samples, strict criteria had to be
observed with these conditions. The addition of ammonium acetate (Beck et al., 2005; met: the chromatographic retention time of the analyte in the sample could not vary
Erbs et al., 2007) and acid (Bacaloni et al., 2005; Kang et al., 2006; Laganà et al., 2004) to more than 2%, and the relative abundance of the two SRM transitions monitored had to
the mobile phase has been examined for their applicability. Thus, the choice of adding lie within the 25% margin, compared to the calibration standards.
0.1% of formic acid to the mobile phase enabled the detection of the phytoestrogens and
also increased the response of the progestogens which are acquired in the same
chromatographic run. 3.3. Results of the analysis

3.2. Method performance
The analytical method developed showed satisfactory performance characteristics
in terms of linearity, repeatability, accuracy, selectivity and sensitivity. Quantitation was
performed by external standard calibration. Five-to-nine point calibration curves, based
on peak areas, were calculated for each of the two SRM transitions selected per
compound from the LC-ESI-MS/MS analysis of standard solutions at concentration
ranges between approximately the limit of quantitation and 250 ng/L (500 ng/L for DAI
and GEN). Calibration curves were shown to be linear, with correlation coefficients (r2)
higher than 0.99, for all compounds except levonorgestrel (see Table 3).
The accuracy and precision of the method were estimated from the replicate
(n = 4) analysis of spiked HPLC grade water with the standard mixture of the analytes
at 100 ng/L. Method extraction recoveries, were above 80% for all compounds, with
exception of DES (see Table 3). Unexpectedly high percentages were obtained for E2S
and DAI, however without any plausible explanation. Test performed with river water
showed lower recovery of the more polar conjugated estrogens, estriol and resveratrol
by a factor higher than 2.5 and up to 8.2. The dilution of the extract by 1/4 could
compensate the matrix load. Hence, it seems that the cause of the obtained lower values
is suppression during ionization prior to MS/MS detection and not a reduced extraction
efficiency due to matrix components. The extracts of the real samples have therefore
been analysed twice to evaluate the extent/importance of this finding.
The repeatability of the procedure was good, with average relative standard
deviation of 7.9. The highest deviations were obtained for genistein and biochanin A
with values of 15.11% and 11.48%, respectively (see Table 3).
A summary of the results, detected concentration range, average values and the
number of positive samples from the target analysis by LC–MS/MS is presented in Table 4.
Eight out of nineteen compounds were found to be present in the studied samples. The
estrogens with the highest estrogenic potency, E2, EE2, DES and E1 were not detected or
not confirmed.
The most ubiquitous compounds, detected in 50% or more of the samples, were
E2G, E3 and PROG (see Table 4). The levels of the above mentioned compounds are also
higher than those of the other estrogens and progestogens. However, these values are in
the range of what is commonly reported, with exception of the 47 ng/L PROG found in
sample 3-R. On the other hand, worryingly high concentrations have been detected of
DAI (276 ng/L), COU (170 ng/L) and GEN (366 ng/L). These values are even higher than
previously described in Osaka Japan (143 ng/L for GEN and 43 ng/L for DAI).
Furthermore, to the authors' knowledge, such high concentration of COU has not been
reported before in river water. The results indicate that these naturally occurring
phytoestrogens have an important urban source (probably due to use for medicinal
purpose), which is subjected to variations according to irregular urban discharges in the
same way as the estrogens and progestogens are.
The quantification of EE2 and NOR was complicated by the presence of interfering
responses. The two SRM transitions from NOR were detected in all samples at the
expected retention time in the chromatogram, however, the ratio of the two transitions
was not concordantly. The average ratio, 299 N145: 299 N171 found was 23.6, instead of
the standard 1.4. The concentrations in the samples calculated with the lowest response
(299 N171) were in the range of 5.1 to 306 ng/L. One could assume the transition with
the higher response might be influenced by a co-eluting compound which exhibits the
same fragmentation. In order to clarify this finding, the samples have been analysed

Fig. 4. Profile of the sum of all detected compounds in each sample in ng/L (⁎ not visible is here that with the sum of DAI, 274 ng/L, and GEN, 366 ng/L, the total value reaches 649 ng/L).
1002 M. Kuster et al. / Environment International 35 (2009) 997–1003
with two additional SRM transitions of NOR (299 N131, 299 N 231), with relative
intensities of ca. 65% towards the main transition, in a shorter chromatographic run
starting at 40% ACN going to 80% ACN in 10 min. Finally, no confirmation of the levels
detected by the two main transitions was obtained, leading to the conclusion that NOR
has not been detected in any sample. In the case of EE2, we were confronted with
another problem. A wide peak of ca. 2 min was found to elute earlier but adjacent to
EE2, in some cases interfering with the beginning of the peak and in other cases the EE2
peak was embedded in this impurity. By monitoring two additional SRM transitions
(295 N143, 295 N183) and using the same approach as described above we hoped to
elude this complication, only to find out that all transitions were affected by the same
interference. Therefore the presence of EE2 could not be confirmed but also not ruled
out.
The analysis of the diluted sample extracts did not confirm the suppression
observed earlier. This finding shows that the influence of the matrix during ionization of
the compounds is highly variable due to its properties and cannot be foreseen by
evaluating the effect in a specific sample.
Fig. 4 illustrates the sum of all compounds in each analyzed sample in both
sampling campaigns. The margin of 10 ng/L is exceeded in 6 sites, but 8 samples (2-R, 3-
R, 4-R, 6-R, 7-1-L, 7-2-L). Locations 2 and 3 are situated up-stream of the Paraíba do Sul
river in a densely populated area with large industrial activities (see Fig. 1). Site 4 is less
populated, however, the river Guandú receives effluents from the waste water and
drinking water treatment plants. Locations 6 and 7 are situated in the metropole of Rio
de Janeiro. Similar or higher values as in sites 2 and 3 were expected in site 5, the
somewhat lower values are probably due to dilution, since the Paraíba do Sul river
receives two important affluent rivers before reaching this site. Even in sample 8-R,
which was considered as a blank river sample, the sum of detected compounds reached
8 and 10 ng/L (see Fig. 4). Summarizing, the locations 1 to 3, 6 and 7 represent the
highest population density and at the same time the highest purchasing power of the
state, also leading to higher consumption of the studied compounds. The levels could
therefore be associated to urban pollution.
Regarding the environmental impact of the reported levels in this study further
research with bio-assays and aquatic organisms have to be performed. Indicatively one
can conclude, based on the potential activity of the conjugated derivatives which might
be cleaved in the environment and the estrogenic activity of the phytoestrogens, free
estrogens and progestogens, that where the total levels pass the 10 ng/L margin a
potential risk of endocrine disrupting effects exists.
4. Conclusions
The multianalyte method developed for the simultaneous extrac-
tion of selected estrogens, progestogens and phytoestrogens and
detection by LC–MS/MS showed overall good performance. Estrogens
were found to be present in concentrations of up to 7.3 ng/L,
progestogens up to 47.1 ng/L, and phytoestrogens up to 366 ng/L. The
results show that most compounds are occasionally present at high
levels, indicating that the rivers studied receive uncontrolled loads of
waste water of different sources and/or that these compounds are not
efficiently removed in the WWTPs discharging into the studied water
bodies. The general ecological status of those rivers is considered to be
poor. The presence of important levels of estrogens, phytoestrogens
and progestogens in river and lagoon water of the studied area can be
a potential risk for the environment and in the future may be for
humans, especially because synergistic and long-term effects are not
fully understood. A general overview concerning this type of urban
pollution was presented in this study. In order to obtain a more
specific assessment of the current and future situation further studies
have to be performed over a longer period of time.
Acknowledgements
This work has been supported by FUJB, CNPq, FAPERJ, the EU
project MODELKEY [GOCE 511237] and by the Spanish Ministry of
Education and Science through the project CEMAGUA (CGL2007-
64551/HID) and reflects the author's view. The EU is not liable for any
use that may be made of the information contained in it. Marina
Kuster gratefully acknowledges the I3P Program (Itinerario integrado
de inserción profesional; co-financed by CSIC and European Social
Funds) for a pre-doctoral grant. Waters is acknowledged for the gift of
SPE cartridges and Merck for the gift of both SPE cartridges and LC
columns.
References
Bacaloni A, Cavaliere C, Faberi A, Foglia P, Samperi R, Lagana A. Determination of
isoflavones and coumestrol in river water and domestic wastewater sewage
treatment plants. Anal Chim Acta 2005;531:229–37.
Baronti C, Curini R, D'Ascenzo G, Di Corcia A, Gentili A, Samperi R. Monitoring natural
and synthetic estrogens at activated sludge sewage treatment plants and in a
receiving river water. Environ Sci Technol 2000;34:5059–66.
Beck I-C, Bruhn R, Gandrass J, Ruck W. Liquid chromatography–tandem mass
spectrometry analysis of estrogenic compounds in coastal surface water of the
Baltic Sea. J Chromatogr A 2005;1090:98-106.
Belfroid AC, Van der Horst A, Vethaak AD, Schafer AJ, Rijs GBJ, Wegener J, et al. Analysis
and occurrence of estrogenic hormones and their glucuronides in surface water and
waste water in The Netherlands. Sci Total Environ 1999;225:101–8.
Bovee TFH, Helsdingen RJR, Rietjens IMCM, Keijer J, Hoogenboom RLAP. Rapid yeast
estrogen bioassays stably expressing human estrogen receptors [alpha] and [beta],
and green fluorescent protein: a comparison of different compounds with both
receptor types. J Steroid Biochem Mol Biol 2004;91:99-109.
Chen C-Y, Wen T-Y, Wang G-S, Cheng H-W, Lin Y-H, Lien G-W. Determining estrogenic
steroids in Taipei waters and removal in drinking water treatment using high-flow
solid-phase extraction and liquid chromatography/tandem mass spectrometry. Sci
Total Environ 2007;378:352–65.
Clotfelter ED, Rodriguez AC. Behavioral changes in fish exposed to phytoestrogens.
Environ Pollut 2006;144:833–9.
Desbrow C, Routledge EJ, Brighty GC, Sumpter JP, Waldock M. Identification of
estrogenic chemicals in STW effluent. 1. Chemical fractionation and in vitro
biological screening. Environ Sci Technol 1998;32:1549–65.
Erbs M, Hoerger CC, Hartmann N, Bucheli TD. Quantification of six phytoestrogens at the
nanogram per liter level in aqueous environmental samples using 13C3-labeled
internal standards. J Agric Food Chem 2007;55:8339–45.
Gentili A, Perret D, Marchese S, Mastropasqua R, Curini R, Di Corcia A. Analysis of free
estrogens and their conjugates in sewage and river waters by solid-phase extraction
then liquid chromatography–electrospray-tandem mass spectrometry. Chromato-
graphia 2002;56:25–32.
Halling-Sørensen B, Nors Nielsen S, Lanzky PF, Ingerslev F, Holten Lützhøft HC,
Jørgensen SE. Occurrence, fate and effects of pharmaceutical substances in the
environment — a review. Chemosphere 1998;36:357–93.
Hanselman TA, Graetz DA, Wilkie AC. Manure-borne estrogens as potential environ-
mental contaminants: a review. Environ Sci Technol 2003;37:5471–8.
Hansen P-D, Dizer H, Hock B, Marx A, Sherry J, McMaster M, et al. Vitellogenin — a
biomarker for endocrine disruptors. Trends Anal Chem 1998;17:448–51.
Hwang CS, Kwak HS, Lim HJ, Lee SH, Kang YS, Choe TB, et al. Isoflavone metabolites and
their in vitro dual functions: they can act as an estrogenic agonist or antagonist
depending on the estrogen concentration. J Steroid Biochem Mol Biol 2006;101:
246–53.
Isobe T, Shiraishi H, Yasuda M, Shinoda A, Suzuki H, Morita M. Determination of
estrogens and their conjugates in water using solid-phase extraction followed by
liquid chromatography–tandem mass spectrometry. J Chromatogr A 2003;984:
195–202.
Kang J, Price WE, Hick LA. Simultaneous determination of isoflavones and lignans at
trace levels in natural waters and wastewater samples using liquid chromato-
graphy/electrospray ionization ion trap mass spectrometry. Rapid Commun Mass
Spectrom 2006;20:2411–8.
Kawaguchi M, Ishii Y, Sakui N, Okanouchi N, Ito R, Inoue K, et al. Stir bar sorptive
extraction with in situ derivatization and thermal desorption-gas chromatography–
mass spectrometry in the multi-shot mode for determination of estrogens in river
water samples. J Chromatogr A 2004;1049:1–8.
Kawanishi M, Takamura-Enya T, Ermawati R, Shimohara C, Sakamoto M, Matsukawa K,
et al. Detection of genistein as an estrogenic contaminant of river water in Osaka.
Environ Sci Technol 2004;38:6424–9.
Koh YKK, Chiu TY, Boobis A, Cartmell E, Lester JN, Scrimshaw MD. Determination of
steroid estrogens in wastewater by high performance liquid chromatography–
tandem mass spectrometry. J Chromatogr A 2007;1173:81–7.
Kuch HM, Ballschmiter K. Determination of endocrine-disrupting phenolic compounds
and estrogens in surface and drinking water by HRGC-(NCI)-MS in the picogram
per liter range. Environ Sci Technol 2001;35:3201–6.
Kuster M, Lopez de Alda MJ, Rodriguez-Mozaz S, Barceló D. Analysis of steroid estrogens
in the environment. In: Petrovic M, Barceló D, editors. Comprehensive Analytical
Chemistry. Elsevier; 2007.
Kuster M, López de Alda MJ, Hernando MD, Petrovic M, Martín-Alonso J, Barceló D.
Analysis and occurrence of pharmaceuticals, estrogens, progestogens and polar
pesticides in sewage treatment plant effluents, river water and drinking water in
the Llobregat river basin (Barcelona, Spain). J Hydrol 2008;358:112–23.
Laganà A, Bacaloni A, De Leva I, Faberi A, Fago G, Marino A. Analytical methodologies for
determining the occurrence of endocrine disrupting chemicals in sewage treatment
plants and natural waters. Anal Chim Acta 2004;501:79–88.
Lopez de Alda MJ, Diaz-Cruz S, Petrovic M, Barcelo D. Liquid chromatography–(tandem)
mass spectrometry of selected emerging pollutants (steroid sex hormones, drugs and
alkylphenolic surfactants) in the aquatic environment. J Chromatogr A 2003;1000:
503–26.
Petrovic M, Sole M, de Alda MJL, Barcelo D. Endocrine disruptors in sewage treatment
plants, receiving river waters, and sediments: integration of chemical analysis and
biological effects on feral carp. Environ Toxicol Chem 2002;21:2146–56.
Reddy S, Iden CR, Brownawell BJ. Analysis of steroid conjugates in sewage influent and
effluent by liquid chromatography–tandem mass spectrometry. Anal Chem
2005;77:7032–8.
 M. Kuster et al. / Environment International 35 (2009) 997–1003
Reemtsma T. The use of liquid chromatography–atmospheric pressure ionization-mass
spectrometry in water analysis — Part II: Obstacles. TrAC, Trends Anal Chem 2001;20:
533–42.
Rodriguez-Mozaz S, López de Alda MJ, Barceló D. Monitoring of estrogens, pesticides
and bisphenol A in natural waters and drinking water treatment plants by solid-
phase extraction-liquid chromatography–mass spectrometry. J Chromatogr A
2004;1045:85–92.
Routledge EJ, Sheahan D, Desbrow C, Brighty GC, Waldock M, Sumpter JP. Identification
of estrogenic chemicals in STW effluent. 2. In vivo responses in trout and roach.
Environ Sci Technol 1998;32:1559–65.
1003
Salste L, Leskinen P, Virta M, Kronberg L. Determination of estrogens and estrogenic
activity in wastewater effluent by chemical analysis and the bioluminescent yeast
assay. Sci Total Environ 2007;378:343–51.
Xiao X-Y, McCalley DV, McEvoy J. Analysis of estrogens in river water and effluents using
solid-phase extraction and gas chromatography-negative chemical ionisation mass
spectrometry of the pentafluorobenzoyl derivatives. J Chromatogr A 2001;923:
195–204.