You are on page 1of 10

Identification of cocaine and its metabolites in urban

wastewater and comparison with the human excretion profile
in urine
Sara Castiglioni
*, Renzo Bagnati
, Manuela Melis
, Deepika Panawennage
Paul Chiarelli
, Roberto Fanelli
, Ettore Zuccato
Department of Environmental Health Sciences, Mario Negri Institute for Pharmacological Research, via La Masa 19, 20156 Milan, Italy
Department of Chemistry, Loyola University, 1032 W. Sheridan Rd., Chicago, IL 60660, USA
a r t i c l e i n f o
Article history:
Received 10 February 2011
Received in revised form
5 July 2011
Accepted 13 July 2011
Available online 23 July 2011
Urban wastewater
Mass spectrometry
Human excretion profile
Sewage epidemiology
a b s t r a c t
The most relevant human urinary metabolites of cocaine (nine metabolites) were
measured in urban wastewater in Italy and USA. A novel analytical method based on liquid
chromatography tandem mass spectrometry allowed the identification of ecgonine,
ecgonine methyl ester and the pyrolytic derivatives of cocaine in untreated wastewater.
The aim of this study was to verify whether the pattern of cocaine metabolites in waste-
water reflected the human excretion profile in urine. The performance of the method was
good, with recoveries higher than 60% and limits of quantifications in the low ng/L range.
The stability in untreated wastewater was assessed for all metabolites and the best storage
condition resulted freezing samples immediately after collection and keep them frozen
until analysis. All the selected compounds were measured in wastewater at concentrations
up to 1.5 mg/L and their weekly loads were calculated during a five weeks monitoring
campaign in Milan (Italy). The profiles of cocaine metabolites in wastewater matched with
those in human urine reported in the literature, suggesting that measures in wastewater
reflect the real human excretion and that wastewater analysis is suitable for assessing drug
consumption. Benzoylecgonine was confirmed as the best target for estimating cocaine use
by wastewater analysis, while cocaine itself should not be considered because its amount
in wastewater is affected by other environmental sources such as transport, handling and
consumption. Results suggested that the measurement of other metabolites in combina-
tion with benzoylecgonine might reflect 60% of an administered dose of cocaine providing
also information on different patterns of use.
ª 2011 Elsevier Ltd. All rights reserved.
1. Introduction
In the last few years illicit drugs have been “discovered” as
a novel class of environmental contaminants, and have
attracted interest in different scientific disciplines such as
analytical and environmental chemistry (Richardson, 2009;
Zuccato and Castiglioni, 2009) and social sciences (EMCDDA,
2008; Frost et al., 2008). The main source of environmental
contamination by illicit drugs is human consumption. In fact,
after the ingestion of a “drug dose”, active parent compounds
(unchanged parent drugs) or metabolites are excreted in
consumers’ urine, entering urban wastewater and reaching
sewage treatment plants (STPs). These substances can persist
even in treated wastewater, and make their way into surface
* Corresponding author. Tel.: þ39 02 39014776; fax: þ39 02 39014735.
E-mail address: (S. Castiglioni).
Avai l abl e at www. sci encedi r ect . com
j our nal homepage: www. el sevi er . com/ l ocat e/ wat r es
wa t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 5 1 4 1 e5 1 5 0
0043-1354/$ e see front matter ª 2011 Elsevier Ltd. All rights reserved.
water (rivers, lakes, sea) or drinking water (Chiaia et al., 2008;
Huerta-Fontela et al., 2008; Jones-Lepp et al., 2004; Postigo
et al., 2008; Zuccato and Castiglioni, 2009).
The first multiresidue analytical method to detect illicit
drugs in urban untreated wastewater was set up by our group
in 2006 (Castiglioni et al., 2006), and was employed to measure
the most widely used classes of illicit drugs, i.e. cocaine and its
main metabolites, amphetamines, opioids, and cannabinoids,
in Italy and Switzerland. The method was based on solid
phase extraction (SPE) and liquid chromatography tandem
mass spectrometry (LC-MS-MS), which became the most
common techniques to detect drugs of abuse in environ-
mental samples (Castiglioni et al., 2008; Postigo et al., 2008).
An interesting application of the measurement of illicit
drugs in wastewater is “sewage epidemiology”, a novel
approachfor estimating drug consumption in a community by
wastewater analysis (Daughton, 2011; Zuccato et al., 2005,
2008). This method is based on the principle that if drug
residues are excreted by humans in substantial amounts and
are sufficiently stable in wastewater, they can be easily
detected in urban untreated wastewater, reflecting the
amounts consumed by the population. Wastewater analysis
has been used to estimate drug consumption in several case
studies (van Nuijs et al., 2010), and drug consumption profiles
obtained from wastewater analysis resulted similar to the
national consumption profiles based on annual prevalence
data. The correspondence between the pattern of drug
metabolites detected in wastewater and their human excre-
tion profile is one of the main assumptions on which the
present method is based (Zuccato et al., 2008), but this corre-
spondence has only been assessed for a few compounds so
far. Moreover, a proper storage of samples is mandatory in
order to avoid the degradation of the analytes that should be
used for drug consumption estimation.
The aim of this study was to verify whether the pattern
of cocaine metabolites detected in wastewater reflects the
human excretion profile in urine. We chose cocaine because
this substance has a well known excretion profile in urine
(Baselt, 2004) with specific metabolites for the different
routes of administration. The complete set of cocaine
metabolites normally detected in consumer’s urine was
measured in untreated wastewater extending and revising
our previous publications (Castiglioni et al., 2008). In
particular, ecgonine is hard to extract from biological fluids
such as plasma (Klingmann et al., 2001), urine or whole
blood (Paul et al., 2005), because of its high polarity and
water-solubility, and required a specific procedure for
extraction from wastewater too. The novel method we have
developed within this study allows the simultaneous anal-
ysis of ecgonine and other structurally-related metabolites
of cocaine, i.e. ecgonine methyl ester, and the main pyro-
lytic products of cocaine (anhydroecgonine and anhy-
droecgonine methyl ester), in untreated wastewater.
Hydrophilic interaction chromatography (HILIC) has been
employed for these substances since it is a powerful tech-
nique for the retention of polar analytes offering a differ-
ence in selectivity compared to traditional reversed-phase
chromatography (Hsieh, 2008). A similar method based on
SPE and direct injection liquid chromatography was devel-
oped contemporaneously to measure these compounds in
wastewater (Bisceglia et al., 2010), but a lower method
performance for ecgonine has been obtained.
The selected compounds were measured in several STPs in
Italy and USA, allowing the comparison of their levels in
wastewater with those measured in urine (Paul et al., 2005)
and with the human excretion profiles of cocaine reported in
the literature. Measuring cocaine metabolites in urban
untreated wastewater was also useful to identify the different
routes of administration of cocaine and to distinguish the
amounts of cocaine fromhumanconsumption and fromother
sources such as production, transport, and distribution.
Finally, the best storage conditions for wastewater samples
has been provided within this study through a stability test
which assessed the degradation of each compound in
untreated wastewater under different storage conditions.
2. Materials and methods
2.1. Selection of compounds
Cocaine is one of the most widely used illicit drugs world-
wide, with between 16 and 21 million of consumers, i.e. 8%
of the total drug users in the population aged 15 to 64 (UNODC,
2009) and has a wide range of human metabolites. Within this
study the compounds selected were: cocaine (COC), benzoy-
lecgonine (BZE), norbenzoylecgonine (NBE), norcocaine (NCO),
cocaethylene (CET), ecgonine methyl ester (EME), ecgonine
(ECG), anhydroecgonine (AEC), anhydroecgonine methyl ester
BZE and EME are the primary human metabolites of COC
and urinary excretion amounts to respectively 45% and 40%
(mean percentages) of the dose consumed (Baselt, 2004;
Ambre et al., 1984; Ambre et al., 1988). ECG is excreted in
smaller amounts in urine, 1e8% of the dose consumed (Fish
and Wilson, 1969; Smith et al., 2010), but can also be
produced by the spontaneous hydrolysis of EME in urine
(Baselt, 2004). COC is only partially excreted in the urine as
unchanged drug, i.e. 1e9% of the administered dose,
depending on the pH of the urine (Baselt, 2004). Other minor
metabolites that formduring the oxidative metabolismof COC
and BZE are NCO, a pharmacologically active metabolite, and
NBE (Maurer et al., 2006). In the presence of ethanol, COC can
be trans-esterified, forming CET, which accounts for 0.7% of
a dose in 24-h urine (Baselt, 2004). When COC is smoked as
crack, i.e. the free base form of COC, the main pyrolytic
product is AME, which is inhaled and form three metabolites
in human body: AEC, anhydroecgonine ethyl ester, and nor-
anhydroecgonine (Maurer et al., 2006). Our investigation
included the main pyrolytic derivative of COC, AME, and one
of its metabolites, AEC, that were detected in human urine
(Paul et al., 2005).
2.2. Chemicals and materials
The analytical standards of COC, BZE, NBE, NCO, CET, EME,
ECG, AEC, AME and the deuterated analogues BZE-d
, COC-d
, ECG-d
, and EME-d
were acquired from Cerilliant
Corporation (Round Rock, Texas, USA). The standards, avail-
able as solutions in methanol or acetonitrile (1 or 0.1 mg/mL),
wa t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 5 1 4 1 e5 1 5 0 5142
were diluted to 10 ng/mL with methanol or acetonitrile, and
stored at À20

C in the dark. Working solutions (1 and 0.1 ng/
mL) containing all the substances to be analyzed were prepared
before each analytical run and stored at À20

C in the dark.
The deuterated compounds were used as internal standards
(IS), and separate working solutions (1 ng/mL) were prepared
before analysis.
Methanol for pesticide analysis, acetonitrile for HPLC-MS
and hydrochloric acid (37%) were from Carlo Erba (Italy);
ammonium hydroxide solution (25%), acetic acid for LC-MS
(>99%), ammonium formate (>99% dry matter) and formic
acid (98%) were acquired from Fluka (Buchs, Switzerland).
HPLC grade MilliQ water was obtained with a MILLI-RO PLUS
90 apparatus (Millipore, Molsheim, France). The cartridges
used for solid phase extraction were 3-mL disposable Oasis
MCX (60 mg), and 6-mL disposable Oasis MCX (150 mg) from
Waters Corp., Milford, MA. The analytical HPLC columns used
were an XTerra MS C18, 100 Â 2.1 mm, 3.5 mm and an X-Bridge
HILIC 100 Â 2.1 mm, 3.5 mm from Waters Corp., Milford, MA.
2.3. Wastewater samples
Composite 24-h water samples, obtained by pooling water
collected by automatic sampling devices and taking 500 mL
from each mixed sample, were collected after the primary
entrance grids from the influents in different STPs in Italy
(Milan, Como and four towns in Sardinia) and USA (Table 1).
Sampling locations were chosen to investigate areas with
different socio-economic patterns. The plants in Italy were
located in the biggest business city in the north of Italy (Milan),
with a high incoming flowof people for work and amusement,
ina smaller but industrializedcity inthe same area (Como) and
in an island (Sardinia). In Sardinia two cities were chosen on
the cost (Cagliari and Olbia) being characterized by a high
touristic influx, andtwo were choseninthe inland(Sassari and
Nuoro). The STP in USA is the Stickney Water Reclamation
Plant which processes wastewater from the south and west
parts of Chicago. For each plant, composite samples were
collected every day for seven consecutive days. The sampling
was repeated for five weeks in Milan, four weeks in Como, two
weeks in Sardinia and one week in Chicago. All the samples
were collected in dry weather (no rainfall) during March-
September 2009 (Italy) and November-December 2009 (USA).
Water samples were frozen and stored in the dark at À20

Before extraction, samples were filtered first on a glass
microfiber filter GF/A 1.6 mm (Whatman, Kent, U.K.) then on
a mixed cellulose membrane filter 0.45 mm (Whatman, Kent,
2.4. Solid phase extraction (SPE)
The procedures for the extraction and analysis of the selected
compounds required their separationinto two mainanalytical
groups (Table 2).
The first analytical group was extracted adapting the
procedure described in a previous publication (Castiglioni
et al., 2006) to an automated SPE system, GX-274 ASPEC (Gil-
son, Middleton, WI, USA). This method was not applicable to
the second analytical group, especially to ECG, due to the low
molecular weight and high water-solubility of this substance
(Castiglioni et al., 2008). ECG extraction was previously per-
formed from biological fluids and wastewater using cation
exchange cartridges (Bisceglia et al., 2010; Klingmann et al.,
2001; Paul et al., 2005) or Hysphere MM anion sorbent
(Jagerdeo et al., 2008), but the recoveries were lower than 40%.
The extraction procedure from untreated wastewater
required an optimization by testing different cartridges and
pH conditions: Oasis MCX (60 mg) at pH 2 and 6.0, Oasis WCX
(60 mg) at pH 4.0 and 8.0, Oasis MAX (60 mg) at pH 8.0, Oasis
HLB at pH 7.0, Evolute ABN at pH 7.0. The recoveries of ECG
were poor in all the cases and the spiked amounts were found
in the wastewater passed through the cartridge, indicating
that ECG was not retained on the solid phases because of its
high water-solubility. The extraction of the other compounds
belonging to the second analytical group was best using an
Oasis MCX cartridge (60 mg) at pH 2.0, so cartridges packed
Table 1 e Main characteristics of the wastewater treatment plants (STPs) investigated.
Mean daily flow
rate (m
by the plant
Type of sewage Wastewater
Milan (Nosedo) Volume proportional 370,000 1,250,000 domestic activated sludge þ disinfection
(peracetic acid)
Como Time proportional
every 20 min
63,000 100,000 domestic þ
activated sludge
Sardinia Time proportional
every 15 min
Cagliari 96,000 300,000 domestic activated sludge þ disinfection
Olbia 20,000 50,000 domestic þ
activated sludge þ disinfection
Sassari 36,000 130,000 domestic þ
activated sludge þ disinfection
Nuoro 7000 21,000 domestic activated sludge þ disinfection
Volume Proportional 1,925,000 2,400,000 domestic þ
activated sludge þ disinfection
wa t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 5 1 4 1 e5 1 5 0 5143
with more sorbent (150 mg) were tested to increase the
adsorption surface. In this way the extraction efficiency of
ECG was highly improved. The volume of extraction was
optimized for the retention of ECG on Oasis MCX (150 mg) by
testing different volumes of untreated wastewater, i.e. 10, 20,
30, 40, and 50 mL. The best result was obtained with an Oasis
MCX cartridge (150 mg, 6 cc) at pH 2.0 and extracting 20 mL of
untreated wastewater. Briefly, untreated wastewaters (20 mL)
were spiked with 20 ng of each IS and the pH was adjusted to
2.0 with 37% HCl. The cartridges were conditioned by washing
with 12 mL methanol and 6 mL MilliQ water, and samples
were passed through by gravity. Cartridges were then
vacuum-dried for 10 min and eluted with 2 mL methanol and
2 mL 2% ammonia solution in methanol. The eluates were
pooled and dried to 10 mL under a nitrogen stream.
2.5. Liquid chromatography (LC) - mass spectrometric
analysis (MS-MS)
The LC system consisted of two Series 200 pumps and a Series
200 auto sampler (PerkineElmer, Norwalk, CT). The first
analytical group (Fig. S-1) was analyzed according to the
method described previously (Castiglioni et al., 2006, 2008).
The second group that includes small, extremely hydrophilic
molecules hardly retained by reversed-phase chromatog-
raphy using acidic buffers, was previously analyzed using two
different chromatographic methods (Castiglioni et al., 2008):
an XTerra C
column with an ammonium formate buffer for
EME and AME, and a PFP Propyl column with an acetic acid
buffer for ECG and AEC. Despite good results for analytical
standards, the chromatographic separation and resolution in
wastewater samples were very poor and the entire procedure
was time-consuming. Therefore, within this study we set out
to find better and faster conditions for the simultaneous
analysis of these compounds. A Luna CN column was initially
tested, but gave the same results as before, and low repro-
ducibility on successive analytical runs. An X-Bridge HILIC
column was then tested, and gave the best results, allowing
simultaneous analysis of all these compounds and giving
good chromatographic resolution both for standards and
wastewater samples. The Hydrophilic Interaction Liquid
Chromatography (HILIC) technique works similarly to
a normal phase chromatography, i.e. the analytes are retained
when gradients are applied with initial high percentage of
organic phase and elution is done with a slightly acidic water
solvent (no more than 50%of water solvent must be applied to
the column). Under these conditions small polar compounds,
like those considered here, can be retained by the stationary
phase, allowing good chromatographic separation. The
eluates concentrated to 10 mL were diluted with 90 mL aceto-
nitrile, then centrifuged and transferred into glass vials. A
gradient elution was used for chromatographic separation,
using ammonium formate 5 mM in water acidified to pH 4
with formic acid as solvent A, and acetonitrile as solvent B at
a flowrate of 300 mL/min. The elution started with 5%of eluent
A and 95% of eluent B, followed by a 12-min linear gradient to
50% of eluent A, a 3-min isocratic elution and a 1-min linear
gradient to 5% of eluent A, which was maintained for 7 min to
equilibrate the column. The injection volume was 10 mL.
Typical chromatograms obtained from the analysis of
untreated wastewater are presented in Fig. 1.
The mass spectrometric analysis was performed using an
API 3000 triple quadrupole, equipped with a turbo ion spray
source (AB - Sciex, Thornhill, Ontario, Canada). The analyses
were done in positive ion mode for all the compounds and the
instrumental conditions optimized for each compound are
summarized in Table S-1. A detailed description of mass
spectrometric analysis, including the collision-induced
dissociation (CID) spectra of all the drugs, quantification
procedures, and matrix effects control are reported elsewhere
(Castiglioni et al., 2006, 2008).
2.6. Performance of the method
The method performance was assessed in untreated waste-
water for all the compounds. Recoveries and repeatability of
the method were assessed by analyzing untreated wastewater
samples in triplicate. Since relevant amounts of illicit drugs
were already present in wastewater, the samples were spiked
with 5 mg/L of BE, 2 mg/L of COC, ECG, EME, AEC, and AME, and
0.5 mg/L of the other drugs before extraction. Blank samples
(mineral water) were analyzed within each analytical run to
test for and correct bias. Instrumental quantification limits
(IQL) were determined by direct injection of standard solu-
tions with increasing amounts of each substance. The limits
of quantification (LOQ) for the whole method were calculated
directly from extracted samples of STP influents. LOQs were
calculated as the concentrations giving peaks with a signal-to-
noise ratio of 10 for the first group of analytes and 5 for the
second group (Table 2). The linearity of the calibration curves
was tested in the ranges normally measured in wastewater
and a calibration curve was injected during each analytical
run to check the linearity (correlation factors) and the
instrumental repeatability. Intra- and interday instrumental
repeatability and precision were also assessed by replicated
injections of standard mixtures and wastewater samples.
2.7. Stability in wastewater
The stability of the selected compounds in untreated waste-
water was investigated to establish the best storage
Table 2 e Recoveries of COC and its main metabolites
(mean ± standard deviation-SD), instrumental
quantification limits (IQL), and limits of quantification
(LOQ) of the analytical method in untreated wastewater.
Chemicals Recovery Æ SD (%) IQL
First analytical group
BZE 107 Æ 9 22 2.1
NBE 85 Æ 5 7.8 0.6
COC 96 Æ 5 23.7 2.1
NCO 112 Æ 7 19.8 2.5
CET 109 Æ 4 16.6 1.3
Second analytical group
ECG 63 Æ 1 60.1 7.2
EME 101 Æ 8 15.5 2.6
AEC 98 Æ 2 11.2 2.2
AME 62 Æ 4 26.5 7.5
wa t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 5 1 4 1 e5 1 5 0 5144
conditions for wastewater samples before and during the
analyses. Stability tests were run in triplicate on samples
stored in glass bottles in the dark for one and three days, and
on other samples processed with a freeze-and-thaw (F/T)
cycle (performed overnight). No pH adjustment was per-
formed during the stability tests to reproduce the real storage
conditions. Untreated wastewater samples were spiked with
5 mg/L of BZE, 2 mg/L of COC, and 1 mg/L of the other drugs.
Three aliquots of wastewater were stored for three days at

C, and analyzed after one and three days. Three identical
aliquots were prepared separately and were subjected
immediately to an F/T cycle. An additional sample without the
drug spiking was used as control for each set of samples and
was processed the same way. The analyses were carried out
immediately after spiking (T
), after one and three days of
incubation at 4

C (T
and T
), and after the F/T cycle (F/T). The
residual amounts of the substances were calculated by sub-
tracting the amounts already present in the control sample.
3. Results and discussion
3.1. Performance of the method
The recoveries obtained in untreated wastewater were higher
than 60% for all the compounds including ECG, with less than
9% variability (Table 2). To our knowledge, this is the best
result obtained for extracting ECG from biological samples
(Klingmann et al., 2001; Paul et al., 2005) and wastewater
(Bisceglia et al., 2010). The instrumental sensitivity was good
and the IQLs ranged from 8 to 60 pg/injected. LOQs in
untreated wastewater ranged between 0.6 and 2.8 ng/L, except
for ECG (7.2 ng/L) and AME (7.5 ng/L) (Table 2). The analytical
response was linear for all the compounds in the range of
concentrations measured in wastewater and the interday
correlation factors (r
) were higher than 0.9996 with standard
deviations (SD) <0.0003 (Table S-2). Instrumental repeatability
was assessed using replicate injections of standard mixtures
and wastewater and was generally lower than 10% (Table S-2).
3.2. Stability in wastewater
The stability of COC and its metabolites in untreated waste-
water was investigated at different storage conditions and the
results are shown in Fig. 2. The first stability test involved the
analysis of triplicate samples after an F/T cycle. The residual
amounts were compared with those measured immediately
after the drug spiking (T
). All the compounds were stable
during this treatment, except for EME and COC which showed
a slight decrease (about 22% less than the initial
The second stability test involved the analysis of triplicate
aliquots of the spiked samples after one (T
) and three days
) of storage at 4

C in the dark (Fig. 2). Again, the residual
amounts were compared with those measured in the samples
immediately after spiking (T
). COC, NCO and EME declined
Fig. 1 e HILIC-LC-MRM chromatograms of ecgonine, ecgonine methyl ester, anhydroecgonine and anhydroecgonine methyl
ester in untreated wastewater. The chromatogram of anhydroecgonine methyl ester refers to a spiked sample (2 mg/L), the
other chromatograms to real samples (80, 150 and 10 ng/L respectively for ecgonine, ecgonine methyl ester and
wa t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 5 1 4 1 e5 1 5 0 5145
substantially after one (À39, À8 and À44% of the initial
concentrations) and three days (À81, À65 and À76%), and
correspondent increases of BZE, NBE and ECG concentrations
were observed after three days (þ21, þ5 and þ26%of the initial
concentrations). The reactions probably occurring in
untreated wastewater were therefore the demethylation of
COC and EME to BE and ECG and the N-demethylation of COC
and BE to NorCOC and NorBE. Further experiments are
required to ascertain the actual transformation processes
occurring during stability tests. After three days’ storage at

C the CET concentration dropped 38% from the initial
concentration, while AEC and AME were quite stable.
These results are in line with the stability tests conducted
by another group for COC and BZE in wastewater at different
storage conditions (Gonzalez-Marino et al., 2010). Our group
also observed decreases in COC, NCO and CET concentrations
in a previous investigation (Castiglioni et al., 2006), but
although the experimental conditions were the same, they
were lower than in this study. Since the only difference was
the wastewater used, we assume that the nature and
composition of the water deeply influences the biological
degradation of these substances, and their stability must
therefore be carefully checked before samples are stored.
The stability of all the compounds we tested was good after
an F/T cycle, but was lowafter one or three days at 4

C (Fig. 2).
We decided therefore to freeze wastewater samples immedi-
ately after collection and keep them frozen (À20

C) until
analysis. Samples were processed immediately after thawing,
avoiding storage at 4

3.3. Occurrence of cocaine and metabolites in untreated
COC and its main metabolites were measured in untreated
wastewater in five medium-sized STPs in Como (Italy),
Cagliari, Olbia, Sassari and Nuoro (Sardinia, Italy), and in two
large plants in Milan (Italy) and Chicago (USA) (Table 3). ECG
and the pyrolytic derivatives of COC (AEC and AME) were
measured in wastewater for the first time within this study.
All the compounds were detected in wastewater, except ECG
and AEC in Sardinia and AME. The total concentrations of COC
and its metabolites (sum of the single concentrations) in
untreated wastewater were 1.3, 0.64, 0.55, 0.61,0.26, 0.28 and
3.1 mg/L respectively in Milan, Como, Cagliari, Olbia, Sassari,
Nuoro and Chicago. The most abundant compound was BZE,
which was measured at concentrations up to 1.5 mg/L in Chi-
cago. COC, EME and ECG reached concentrations in the
hundreds of ng/L range, and the other metabolites were at
lower concentrations. Similar levels of BZE and COC were
reported in untreated wastewater from other European
countries, such as Spain, Belgium, UK, Croatia, The
Netherlands, and Germany (Hogenboom et al., 2009; Postigo
et al., 2008; Terzic et al., 2010; Zuccato and Castiglioni, 2009).
COC, BZE, NCO and NBE were also measured in seven STPs in
USA (Chiaia et al., 2008), and their concentrations were in the
same range as those in Chicago in this study. The ratio of COC
to BZE was about 0.3 in Milan, Como and Sardinia and 0.55 in
Chicago, and was higher than in a previous analytical
campaign in Italy (0.15) (Zuccato et al., 2005), but similar to
those found in several STPs in Belgium (0.4) (van Nuijs et al.,
2009). This is higher than the ratio expected from the
percentages of excretion of these substances in urine reported
in the literature (1e9% of COC and 35e54% of BZE), which
averages 0.1. This suggests a role for other sources of COC
probably related to transport, handling and consumption of
the drug.
The main pyrolytic metabolite of COC, AME, which is
mainly metabolized in the human body to AEC and other
metabolites (Maurer et al., 2006), was never found in waste-
water, but AEC was detected at measurable concentrations in
Milan (35 samples), Como (28 samples) and Chicago (7
samples) (Table 3). The ratio of AEC (a metabolite of COC when
Fig. 2 e Stability of the compounds in wastewater immediately after drug spiking (T
), after a freeze-and-thaw cycle (F/T),
and after one day (T
) and three days (T
) of storage at 4

wa t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 5 1 4 1 e5 1 5 0 5146
smoked as crack) to BZE (the major metabolite of COC) was
three times higher in Chicago than in Milan. This can suggests
that in the time period under analysis, the percentage of COC
used as crack was higher in Chicago than in Milan, in line with
epidemiological investigations indicating a greater prevalence
of crack use in USA than in Italy (EMCDDA, 2009). Even if
further research is needed to confirm these results, the AEC/
BE ratio seemto be useful as a roughestimate of the amount of
COC smoked as crack in a community, and can therefore help
to estimate how much COC is used as crack.
3.4. Weekly loads of cocaine and metabolites in
untreated wastewater in Milan
The daily loads (g/day) of COC and metabolites have been
calculated in a five weeks monitoring campaign conducted in
Milan, multiplying the concentrations of each compound in
untreated wastewater by the mean daily flow rates (Fig. 3A).
Generally, the loads increased during the weekend reaching
a peak on Saturday, when BZE, EME, ECG, and NBE loads were
30% higher than in a weekday. The load of BZE was the
highest, ranging from 250 g/day during the week to 350 g/day
on Saturday. Lower loads were measured for EME and ECG,
and were respectively 60 and 35 g/day during the week and 90
and 45 g/day on Saturday. Among the minor metabolites of
COC, NBE was the most abundant (7e10 g/day) and CET
showed a marked increase during the weekend (from 2.5 to
3.7 g/day). The daily loads of NCO and AEC were about 2 g/day
without any relevant change during the weekend.
Fig. 3B shows the weekly loads of COCcompared to those of
BZE and of the sum of all the measured metabolites (including
BE). The daily load of COC ranged from 90 g/day in a weekday
to 110 g/day on Saturday, thus following the same weekly
pattern of its main metabolites. Despite this suggests a rela-
tionship with human urinary excretion, the total amount of
COC found in wastewater is higher than that expected from
human metabolism, as discussed in the previous paragraph,
and it is therefore unsuitable to indicate human consumption.
The loads of all metabolites were 30%higher than those of BZE
alone (Fig. 3B), indicating that the measurement of additional
COC metabolites in combination with BE could increase the
percentage of an administered dose detected in wastewater. If
BE alone represents about 45% of a dose of COC, all
Table 3 e Mean levels (ng/L) of COC and its main metabolites and Relative Standard Deviation (RSD%) in untreated
wastewater in Italy and USA detected during several successive campaigns carried out in 2009.
Chemicals Milan (five weeks
Como (four weeks
Sardinia (two weeks
Chicago (one
week monitoring)
Cagliari Olbia Sassari Nuoro
Served people 1,250,000 101,000 300,000 50,000 130,000 21,000 2,400,000
BZE 712 (8) 380 (10) 316 (2.5) 337 (14) 149 (11) 153(8) 1553 (22)
NBE 20 (24) 12 (25) 11 (8) 11 (37) 5.5 (11) 5.3 (6) 51 (20)
COC 255 (15) 98 (38) 138 (37) 131 (5) 48 (13) 55 (16) 868 (21)
NCO 4.0 (38) 3.2(12) 3.5 (33) 3.3 (29) 0.9 (7) 2.0 (24) 28 (23)
CET 6.4 (16) 2.7(29) 3.0 (9) 5.1 (33) 0.4 (10) 3.2 (14) 23 (40)
ECG 97 (8) 58 (35) <LOQ <LOQ <LOQ <LOQ 199 (23)
EME 176 (11) 84 (29) 96 (9) 119 (27) 58 (11) 68 (9) 346 (23)
AEC 4.3 (29) 1.0 (23) <LOQ <LOQ <LOQ <LOQ 35 (27)
Total amount
1276 (18) 638 (20) 548 (21) 607 (20) 257 (20) 285 (21) 3103 (17)
Fig. 3 e A. Daily loads (g/day) of COC and metabolites in
untreated wastewater in Milan. B. Comparison between
the daily loads of COC, and BZE and the sum of all
metabolites measured in wastewater. The loads are mean
values (with standard deviation) of a five week monitoring
wa t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 5 1 4 1 e5 1 5 0 5147
metabolites can reflect 59e60%. Moreover, the analysis of
different COC metabolites in wastewater might also help
investigating the patterns of consumption. For instance,
measuring CET might help to assess how frequently COC is
consumed together with alcohol as observed for its weekly
pattern of consumption in Milan (Fig. 3A).
3.5. Comparison of metabolic profiles in wastewater and
in human urine
The profile of COC metabolites in wastewater was compared
with the profiles in human urine reported in the literature.
The urinary profiles considered came either from pharmaco-
kinetic studies conducted under controlled dose-
administration conditions, and from spot urine analysis
(Table 4). Table 4 (first column) reports the amount of each
metabolite, as a percentage of the administered dose of COC,
recovered in the 24e72 h urine in pharmacokinetic studies.
The main limitation of these data is the relatively low number
of subjects involved (2e6 subjects each study). Thus we
decided to include in our comparison supplementary data
from spot urine analysis (Table 4, second column). Each
metabolite is reported as a percentage of the sum of the
median levels of each metabolites plus COC measured in spot
urine specimens. These percentages were calculated from
data published recently by Paul et al., who measured 15 COC-
related compounds in 30 urine samples from living individ-
uals and post-mortem specimens. The excretion profiles of
COC metabolites and COC were calculated using median
values because of the high variability of the concentrations
measured in spot urine samples (Paul et al., 2005). Similarly,
the third column reports each metabolite as a percentage of
the sum of the median concentrations of metabolites plus
COC measured in wastewater. These percentages were
calculated using the concentrations measured in Milan and
Como (Italy), and Chicago (USA) because a similar pattern of
COC consumption has been found in these cities.
Due to the variable excretion ranges observed for some
compounds in pharmacokinetics studies (i.e. cocaine, range
1e14%), only a general comparison among the excretion
profiles was possible according to the aims of our study. Data
from wastewater could be reasonably compared with profiles
in urine (Table 4), since they indicate the profiles of excretion
of the different metabolites of COC in the urine of single
individuals (pharmacokinetic studies), and in wastewater,
which collects the urine excreted by an entire community.
BZE was confirmed as the most abundant metabolic product
of COC, with excretion approximating 50% in all cases.
Percentages were comparable also for other minor metabo-
lites such as NBE, NCO and CET. On the contrary, some
differences were observed for EME, which is indicated as the
second main urinary metabolite of COC with 32e49% excre-
tion in pharmacokinetic studies, and 17% in spot urine
analysis, but was detected in wastewater at a lower
percentage (12.7%), probably because of its limited stability in
wastewater, as also assessed in our study (Fig. 2). The
percentage of ECG found in wastewater (8%) was comparable
with the figures from pharmacokinetic studies (1e8%), but
lower than in spot urine analysis (21%). Since ECG is indi-
cated as the main transformation product of EME during the
Table 4 e Comparison of the metabolic profiles of COC in human urine (from pharmacokinetic studies and spot urine
analysis) and wastewater.
Substances Percentages of COC
and its metabolites from
pharmacokinetic studies
(2e6 subjects)
Percentages of COC
and its metabolites from
spot urine analysis
(30 subjects)
Percentages of COC
and its metabolites
from wastewater
analysis (entire community)
BZE 35e54
49 54.8
NBE 3.1
2.5 1.7
COC 1e9
3.7 21.2
NCO e 0.6 0.5
CET 0.7
1.5 0.6
ECG 2-8
21 8.3
EME 32e49
17 12.7
AEC e 4.4 0.5
AME 0.02
0.3 e
Fish and Wilson, 1969;
Ambre et al., 1984;
Baselt, 2004;
Hamilton et al., 1977;
Huestis et al., 2007;
Smith et al., 2010;
Cone et al., 1998.
a Mean percentages calculated from the reported C
following controlled oral, intravenous, intranasal and smoked administration.
wa t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 5 1 4 1 e5 1 5 0 5148
storage of urine samples (Klingmann et al., 2001), a delay
between urine generation and sampling might explain this
relatively high percentage of ECG and the low percentage of
EME in spot urine. The percentage of COC measured in
wastewater (21% of the total) was much higher than expected
from pharmacokinetic studies (1e14% of the administered
dose) or spot urine analysis (3.7% of the total). These results
suggest COC might enter wastewater from other environ-
mental sources besides human urinary excretion, as indi-
cated by the presence of traces of COC on money (Armenta
and de la Guardia, 2008) and airborne particulate (Cecinato
et al., 2009; Postigo et al., 2009).
The metabolic pattern of COC when smoked as crack
depends on the pyrolytic conditions and composition of the
drug dose (Maurer et al., 2006) and is therefore subjected to
wide variability. The average percentages of these metabolites
in spot urine samples of suspected consumers of COC by an
unknown route were 4.4% and 0.3% of the total metabolites
respectively for AEC and AME (Paul et al., 2005). AEC was lower
in wastewater (0.5% of the total metabolites), as expected,
since only a small proportion of consumers may use COC as
crack, the majority being more likely to “snort” the drug.
4. Conclusions
Profiles of COC metabolites in wastewater qualitatively and
quantitatively matched with the profiles in human urine
suggesting that concentrations of metabolites in wastewater
reflect real humanexcretion and that wastewater analysis can
be used to estimate collective urinary excretion of COC
Studying the stability of cocaine metabolites in wastewater
is essential to figure out eventual losses that can occur in the
sewer before wastewater enter STPs, during sampling and/or
the storage of samples. BZE occurrence and stability in
wastewater make it the best target for estimating community
drug use by wastewater analysis. COC itself should not be
considered for estimating COC consumption, as its amount in
wastewater is probably affected by other environmental
sources besides human metabolism, and this might cause
overestimation of the real amounts consumed. Indications
about COC consumption and pattern of use can also be high-
lighted by measuring metabolites other than BZE in waste-
water. However, considering the documented inter-
conversion of some metabolites, an extended panel of
substances should be used, including all the metabolites
measured in this study. In this case, the bulk of metabolites
found in wastewater would reflect 59e60% of the adminis-
tered dose and could also indicate if cocaine is smoked,
snorted, and consumed together with alcohol.
The novel approach based on wastewater analysis, that we
call “sewage epidemiology”, is therefore suitable for assessing
drug consumption in a population, providing real-time
updates of the patterns of drug consumption, and preserving
individual anonymity. It is also useful to identify different
patterns of consumption in cities and countries and has
several fields of application to complement epidemiological
The authors are grateful to Milano Depur Spa, Como Depur
Spa, Abbanoa Spa and the Metropolitan Water Reclamation
District of Chicago (MWRDC) for assistance and collaboration
for sampling respectively in Milan, Como, Sardinia and Chi-
cago. The authors also thank the Como Local Health Agency
and the Cagliari Local HealthAgency (ASL 8) for supporting the
investigations in Como and Sardinia.
Appendix. Supplementary data
Supplementary data related to this article can be found online
at doi:10.1016/j.watres.2011.07.017.
r e f e r e n c e s
Ambre, J., Fischman, M., Ruo, T.I., 1984. Urinary excretion of
ecgonine methyl ester, a major metabolite of cocaine in
humans. J. Anal. Toxicol. 8, 23e25.
Ambre, J., Ruo, T.I., Nelson, J., Belknap, S., 1988. Urinary excretion
of cocaine, benzoylecgonine, and ecgonine methyl ester in
humans. J. Anal. Toxicol. 12, 301e306.
Armenta, S., de la Guardia, M., 2008. Analytical methods to
determine cocaine contamination of banknotes from around
the world. TrAC - Trends Anal. Chem. 27, 344e351.
Baselt, R.C., 2004. Disposition of Toxic Drugs and Chemicals in
Man. Biomedical Publications, Foster City, California, USA.
Bisceglia, K.J., Roberts, A.L., Schantz, M.M., Lippa, K.A., 2010.
Quantification of drugs of abuse in municipal wastewater via
SPE and direct injection liquid chromatography mass
spectrometry. Anal. Bioanal. Chem. 398, 2701e2712.
Castiglioni, S., Zuccato, E., Chiabrando, C., Fanelli, R., Bagnati, R.,
2008. Mass spectrometric analysis of illicit drugs in wastewater
and surface water. Mass Spectrom. Rev. 27, 378e394.
Castiglioni, S., Zuccato, E., Crisci, E., Chiabrando, C., Fanelli, R.,
Bagnati, R., 2006. Identification and measurement of illicit
drugs and their metabolites in urban wastewater by liquid
chromatography-tandem mass spectrometry. Anal. Chem. 78,
Cecinato, A., Balducci, C., Nervegna, G., 2009. Occurrence of
cocaine in the air of the World’s cities. An emerging problem?
A new tool to investigate the social incidence of drugs? Sci.
Total Environ. 407, 1683e1690.
Chiaia, A.C., Banta-Green, C., Field, J., 2008. Eliminating solid
phase extraction with large-volume injection LC/MS/MS:
analysis of illicit and legal drugs and human urine indicators
in U.S. wastewaters. Environ. Sci. Technol. 42, 8841e8848.
Cone, E.J., Tsadik, A., Oyler, J., Darwin, W.D., 1998. Cocaine
metabolism and urinary excretion after different routes of
administration. Ther. Drug Monit. 20, 556e560.
Daughton, C.G., 2011. Illicit drugs: contaminants in the
environment and utility in forensic epidemiology. Rev.
Environ. Contam. Toxicol. 210, 59e110.
EMCDDA (European Monitoring Centre for Drugs and Drug
Addiction), 2008. Insights 9.
EMCDDA (European Monitoring Centre for Drugs and Drug
Addiction), 2009. Cocaine and crack cocaine. In: Drug Situation
in Europe.
Fish, F., Wilson, W.D., 1969. Excretion of cocaine and its
metabolites in man. J. Pharm. Pharmacol. 21, 135e138S.
wa t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 5 1 4 1 e5 1 5 0 5149
Frost, N., Griffiths, P., Fanelli, R., 2008. Peering into dirty waters:
the potential and implications of a new approach to
monitoring drug consumption. Addiction 103, 1239e1241.
Gonzalez-Marino, I., Quintana, J.B., Rodriguez, I., Cela, R., 2010.
Determination of drugs of abuse in water by solid phase
extraction, derivatization and gas chromatography-ion trap-
tandem mass spectrometry. J. Chrom. A 1217, 1748e1760.
Hamilton, H.E., Wallace, J.E., Shimek, E.L., Land, P., Harris, S.C.,
Christenson, J.G., 1977. Cocaine and benzoylecgonine
excretion in humans. J. Forensic Sci. 22, 697e707.
Hogenboom, A.C., van Leerdam, J.A., de Voogt, P., 2009. Accurate
mass screening and identification of emerging contaminants
in environmental samples by liquid chromatography-hybrid
linear ion trap orbitrap mass spectrometry. J. Chromatogr. A
1216, 510e519.
Hsieh, Y., 2008. Potential of HILIC-MS in quantitative bioanalysis
of drugs and drug metabolites. J. Sep. Sci. 31, 1481e1491.
Huerta-Fontela, M., Galceran, M.T., Ventura, F., 2008. Stimulatory
drugs of abuse in surface waters and their removal in
a conventional drinking water treatment plant. Environ. Sci.
Technol. 42, 6809e6816.
Huestis, M.A., Darwin, W.D., Shimomura, E., Lalani, S.A.,
Trinidad, D.V., Jenkins, A.J., Cone, E.J., Jacobs, A.J., Smith, M.L.,
Paul, B.D., 2007. Cocaine and metabolites urinary excretion
after controlled smoked administration. J. Anal. Toxicol. 3,
Jagerdeo, E., Montgomery, M.A., Lebeau, M.A., Sibum, M., 2008. An
automated SPE/LC/MS/MS method for the analysis of cocaine
and metabolites in whole blood. J. Chromatogr. B. Anal.
Technol. Biomed. Life Sci. 874, 15e20.
Jones-Lepp, T.L., Alvarez, D.A., Petty, J.D., Huckins, J.N., 2004.
Polar organic chemical integrative sampling and liquid
chromatography-electrospray/ion-trap mass spectrometry for
assessing selected prescription and illicit drugs in treated
sewage effluents. Arch. Environ. Contam. Toxicol. 47, 427e439.
Klingmann, A., Skopp, G., Aderjan, R., 2001. Analysis of cocaine,
benzoylecgonine, ecgonine methyl ester, and ecgonine by
high-pressure liquid chromatography-API mass spectrometry
and application to a short-term degradation study of cocaine
in plasma. J. Anal. Toxicol. 25, 425e430.
Maurer, H.H., Sauer, C., Theobald, D.S., 2006. Toxicokinetics of
drugs of abuse: current knowledge of the isoenzymes involved
in the human metabolism of tetrahydrocannabinol, cocaine,
heroin, morphine, and codeine. Ther. Drug Monit. 28, 447e453.
Paul, B.D., Lalani, S., Bosy, T., Jacobs, A.J., Huestis, M.A., 2005.
Concentration profiles of cocaine, pyrolytic methyl ecgonidine
and thirteen metabolites in human blood and urine:
determination by gas chromatography-mass spectrometry.
Biomed. Chromatogr. 19, 677e688.
Postigo, C., Lopez de Alda, M.J., Barcelo´ , D., 2008. Analysis of
drugs of abuse and their human metabolites in water by
LC-MS2: a non-intrusive tool for drug abuse estimation at
the community level. TrAC - Trends Anal. Chem. 27,
Postigo, C., Lopez De Alda, M.J., Viana, M., Querol, X., Alastuey, A.,
Artin˜ ano, B., Barcelo´ , D., 2009. Determination of drugs of abuse
in airborne particles by pressurized liquid extraction and
liquid chromatography-electrospray-tandem mass
spectrometry. Anal. Chem. 81, 4382e4388.
Richardson, S.D., 2009. Water analysis: emerging contaminants
and current Issues. Anal. Chem. 81, 4645e4677.
Smith, M.L., Shimomura, E., Paul, B.D., Cone, E.J., Darwin, W.D.,
Huestis, M.A., 2010. Urinary excretion of ecgonine and five
other cocaine metabolites following controlled oral,
intravenous, intranasal, and smoked administration of
cocaine. J. Anal. Toxicol. 34, 57e63.
Terzic, S., Senta, I., Ahel, M., 2010. Illicit drugs in wastewater of
the city of Zagreb (Croatia)eestimation of drug abuse in
a transition country. Environ. Pollut. 158, 2686e2693.
UNODC (United Nations Office of Drugs and Crime), 2009. World
Drug Report.
van Nuijs, A.L., Castiglioni, S., Tarcomnicu, I., Postigo, C., de
Alda, M.L., Neels, H., Zuccato, E., Barcelo, D., Covaci, A., 2010.
Illicit drug consumption estimations derived from wastewater
analysis: a critical review. Sci. Total Environ.. doi:10.1016/j.
scitotenv.2010.05.030 20598736.
van Nuijs, A.L., Pecceu, B., Theunis, L., Dubois, N., Charlier, C.,
Jorens, P.G., Bervoets, L., Blust, R., Neels, H., Covaci, A., 2009.
Spatial and temporal variations in the occurrence of cocaine
and benzoylecgonine in waste- and surface water from
Belgium and removal during wastewater treatment. Water
Res. 43, 1341e1349.
Zuccato, E., Castiglioni, S., 2009. Illicit drugs in the environment.
Philos. Trans. A Math. Phys. Eng. Sci. 367, 3965e3978.
Zuccato, E., Chiabrando, C., Castiglioni, S., Bagnati, R., Fanelli, R.,
2008. Estimating community drug abuse by wastewater
analysis. Environ. Health Perspect. 116, 1027e1032.
Zuccato, E., Chiabrando, C., Castiglioni, S., Calamari, D.,
Bagnati, R., Schiarea, S., Fanelli, R., 2005. Cocaine in surface
waters: a new evidence-based tool to monitor community
drug abuse. Environ. Health 4, 14.
wa t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 5 1 4 1 e5 1 5 0 5150