Ab Initio and in Situ Comparison of
Caffeine, Triclosan, and Triclocarban
as Indicators of Sewage-Derived
Microbes in Surface Waters
T H A Y E R A . Y O U N G ,†J O C H E N H E I D L E R ,†
C R I S T I N A R . M A T O S - P É R E Z ,†
A M I R S A P K O T A ,†T A N I K K A T O L E R ,†
K R I S T E N E . G I B S O N ,†
K E L L O G G J . S C H W A B ,†A N D
R O L F U . H A L D E N *, † , ‡
Johns Hopkins University Center for Water and Health,
Department of Environmental Health Sciences, Bloomberg
School of Public Health, Johns Hopkins University,
Baltimore, Maryland 21205, and Center for Environmental
Biotechnology, The Biodesign Institute, Arizona State
University, Tempe, Arizona 85287
Received October 13, 2007. Revised manuscript received
February 8, 2008. Accepted February 18, 2008.
in theory and practice for their potential to trace sewage-
hypothesis was that hydrophobic OWCs outperform caffeine
as a chemical tracer, due to their sorptive association
with suspended microorganisms representing particulate
organic carbon (POC). Modeling from first principles (ab initio)
of OWC sorption to POC under environmental conditions
suggested an increasing predictive power: caffeine (0.2%
sorbed) < triclosan (9–60%; pH 6–9) < triclocarban (76%).
Empirical evidence was obtained via analysis of surface water
MD. Mass spectrometric OWC detections were correlated
including multiple chronic sewage release sites and the
local wastewater treatment plant. Consistent with ab initio
calculations, correlation analyses of 104 observations for fecal
coliforms, enterococci, and Escherichia coli in natural
surface waters showed that the particle-active antimicrobials
triclosan and triclocarban (R2range, 0.45–0.55) were indeed
It is concluded that chemical monitoring of microbial risks is
and triclocarban in place of, or in conjunction with, the
traditional marker caffeine.
to find rapid, sensitive, and specific markers which can aid
in the detection and localization of sewerage leaks for
subsequent intervention and mitigation. Potential markers
of wastewater leakage to surface water can be classified as
microbial, biomolecular, or chemical. Bacteria, coliphages,
mammalian viruses, and protozoa typically are enumerated
using culturing techniques, including selective plate counts
(1–3), or biomolecular techniques that target nucleic acids
with the polymerase chain reaction or proteins with enzy-
matic and immunological assays (3–5). Organic wastewater
compounds (OWCs), including fecal steroids, caffeine,
pharmaceuticals, and antimicrobial consumer product ad-
ditives, are chemical markers that can be tracked by gas or
spectrometry (2, 6, 7).
indicators, a major one being the relatively shorter analysis
time. Microbial assays relying on culturing techniques
typically take from 18 to 96 h (8, 9), whereas biomolecular
and chemical analyses can be completed in hours or even
minutes (10, 11). Targeting chemical markers is beneficial
since many microbial indicator organisms extant in raw
sewage also are common to wildlife and sources other than
occur at the highest levels at the source and at relatively
lower concentrations in receiving streams (13, 14), whereas
(15). Due to the large number of microbial pathogens
contained in sewage and their varying culture conditions
and requirements, microbial monitoring typically targets
Escherichia coli (E. coli) (9).
Among the many OWCs employed for source tracking,
caffeine is one of the most frequently used chemical
indicators. Caffeine typically occurs in sewage at levels of
20–300 µg/L (16), and its occurrence in surface waters was
observed to correlate with the presence and abundance of
various microbial contaminants (17). Although the antimi-
crobial compounds triclosan and triclocarban are among
the top 10 OWCs with respect to both occurrence frequency
and concentration (6, 18), their value for microbial source
tracking has not yet been explored in great detail.
When choosing a chemical marker of sewage contamina-
tion in surface waters, it would be desirable to find one that
associates and migrates selectively with microorganisms
sorption to organic particles, as evidenced by a high
octanol–water partitioning coefficient (KOW) and organic
carbon partitioning coefficient (KOC), therefore would be
we evaluated three of the 10 most common OWCs (6, 18),
i.e., caffeine, triclosan, and triclocarban, for their ability to
indicate microbial contaminants in natural waters.
Chemicals. Custom synthesis of triclocarban (TCC; 3,4,4′-
trichlorocarbanilide; CAS No. 101-20-2) uniformly labeled
with carbon-13 in the p-chlorophenyl ring was performed
Ciba Specialty Chemicals (Basel, Switzerland) provided as a
CAS No 3380-34-5) uniformly labeled with carbon-13 in the
2,4-dichlorophenoxy ring (13C6-TCS; 98.7%). Anhydrous caf-
feine (99%; CAS No. 58-08-2) was purchased from Fluka
Chemie GmbH (Buchs, Switzerland). Caffeine labeled with
e-mail: firstname.lastname@example.org; mailing address: 1001 S. McAllister Ave.,
P.O. Box 875701, Tempe, AZ 85287-5701.
†Johns Hopkins University.
‡Arizona State University.
Environ. Sci. Technol. 2008, 42, 3335–3340
10.1021/es702591r CCC: $40.75
Published on Web 04/05/2008
2008 American Chemical SocietyVOL. 42, NO. 9, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY93335
either one or three carbon-13 containing methyl groups (3-
[99 at. %], respectively) was obtained from CDN Isotopes
(Pointe-Claire, Quebec, Canada) or Isotech (Miamisburg,
OH), respectively. Solvents of HPLC grade or purer were
purchased from Fisher Scientific (Pittsburgh, PA). All other
chemicals, including TCC and TCS, were obtained at the
two disposable, precleaned 1 L high-density polyethylene
sample bottles for chemical analyses, and one autoclaved
reusable polypropylene 1 L bottle for bacteriological assays.
Trip blanks consisted of 1 L bottles filled with reagent water
(18.2 MΩ resistance) obtained from a Nanopure Diamond
UV ultrafiltration water system (Barnstead; Dubuque, IA).
All stream samples were collected between October of 2003
and December of 2004. In seven sampling campaigns, river
water was taken at various distances up- and downstream
along 14 streams in three watersheds straddling Baltimore
City and Baltimore County, MD. Sampling sites were dis-
persed among the following streams in each watershed: in
Gwynns Falls watershed (GF), Gwynns Falls, Gwynns Run,
(JF), Jones Falls, Stony Run, and Western Run; in Herring
Run watershed (HR), Armistead Creek, Biddison Run, Chin-
from agricultural, to suburban, to urban/industrial. Sites in
the JF and HR watersheds extend from suburban to urban/
industrial. Grab samples of wastewater influent also were
collected in December of 2005 at the local wastewater
treatment plant (WWTP), Baltimore, MD. Upon arrival in
the laboratory, samples for chemical analysis were fortified
used as described elsewhere (19, 20).
were determined in both the dissolved and adsorbed phase
via solid-phase extraction of the aqueous phase and concur-
rent organic solvent liquid extraction of the pellet obtained
by centrifugation (2000g for 20 min), using protocols
described in detail elsewhere (19, 20). Results are presented
as the sum of the aqueous and solid-phase concentrations.
Chemical Analyses. All samples were analyzed for tri-
clocarban and triclosan using liquid chromatography nega-
tive electrospray ionization quadrupole mass spectrometry
(LC-ESI-MS) and the isotope dilution method as described
previously (18). Caffeine was tracked in positive ionization
mode using polarity switching. The typical retention time
for caffeine was 4.0 min, with characteristic mass-to-charge
ratios of m/z 195, 196, and 198 for nonlabeled,13C1-, and
13C3-labeled caffeine, respectively. To attain enhanced detec-
tion limits, samples from the Herring Run watershed were
mass spectrometry as described in detail elsewhere (19).
regardless of the instrument used. Quality assurance and
quality control protocols for triclosan and triclocarban have
been reported previously (18, 19). Caffeine was recovered
from spiked performance samples at an average recovery
rate of 92 ( 21% over a concentration range of 5-150 µg/L.
consistently yielded nondetect concentrations.
Bacteriological Assays. Bacteria were enumerated using
standard membrane filtration methods. Enterococci and E.
coli were assayed using U.S. EPA Methods 1600 (21) and
1103 (22), respectively. Fecal coliforms were enumerated
using Standard Method SM 9222 D (8).
and the 14 sampled streams were used to predict the
partitioning behavior of each chemical under environmen-
tally relevant conditions. Values for the acid ionization
constant (pKA), the pH-dependent n-octanol–water distribu-
tion ratio (DOW), pH-dependent organic carbon partitioning
coefficient (DOC), and the solubility in water of each com-
pound were extracted from the literature or predicted
(Supporting Information; Table S1). Organic carbon and pH
dependence of partitioning isotherms for three potential
chemical markers of sewage leakage to surface water were
calculated using eq 1,
where fdissolvedis the mass fraction in the aqueous phase, Vw
is the volume of water, KD is the solid-water distribution
coefficient, and MSis the mass of suspended solids. KDcan
be approximated as the product of DOCand fOC, where DOC
described above, and fOCis the fraction of MScomposed of
organic carbon. The product of fOCand MScan be defined
as MPOC, the mass of suspended particulate organic carbon,
to yield eq 2 (23):
Data Analysis. Histogram analysis, before and after log
From this analysis it was determined that log-normal
distribution best described the data. The chemical and
biological data were log-transformed and least-squares
(caffeine, triclosan, triclocarban, and the sum of the latter
indicators (fecal coliforms, enterococci, and E. coli). Cor-
of nondetect values to ensure that similar outcomes result
regardless of the data processing strategy chosen.
Ab Initio Analysis. Physicochemical characteristics of the
three chemical indicators were used to predict the partition-
extant in the 14 urban streams of the three Baltimore
watersheds investigated here (Figure 1). Values for the pKA,
DOC, DOW, and the water solubility of each compound were
taken from the literature or calculated (Supporting Informa-
tion; Table S1). At g3.9 g/L, caffeine was found to have a
hydrophobic antimicrobials triclosan and triclocarban. Caf-
feine and triclocarban were predicted to ionize only in
environmentally atypical conditions of pH < 1 and > 12. In
contrast, due to its pKAvalue of 7.8, triclosan was predicted
protonation (DOCof 15500 at pH 5.4) to almost full depro-
tonation (DOC of 600 at pH 9.2) in the pH range extant in
Baltimore’s streams (pH 5.4–9.2). Monitoring data collected
by the local municipality over the course of 2 years revealed
that, in addition to substantial variations in pH, the inves-
3336 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 9, 2008
streams was estimated from the literature (2.2–35% POC);
raw wastewater entering the local sewage treatment plant
showed an even broader range of 40–363 mg/L TSS and
0.8–41% POC (Supporting Information; Table S2).
These observed ranges in environmental conditions
guided the calculation of pH-dependent partitioning iso-
therms (Figure 2). Caffeine and triclocarban exhibited
constant pH-adjusted KOCvalues (DOC) between pH 6 and 9,
was more complex in that it displayed a pH dependence
characteristic of an acidic hydroxyl group; its mass fraction
6) at a POC concentration of 100 mg/L. Caffeine was
insensitive to the amount of POC present and remained
essentially fully dissolved (99.8%) even at maximum POC
from 100 to 24% as organic carbon levels increased from 0
to 100 mg/L.
Chemical Analyses. Surface waters from all three Balti-
of interest (Figure 3). Caffeine was detected in all streams
but Tiffany Run and peaked at a concentration of 7110 ng/L
FIGURE 1. Map of 34 sampling sites along 14 streams in three watersheds of Baltimore City and County, Maryland. Also shown are
five chronic sources of wastewater leakage to surface water and the wastewater treatment plant (WWTP) monitored in this study.
FIGURE 2. Sorption isotherms illustrating the effect of particulate
organic carbon (POC) and pH on the partitioning behavior of three
organic wastewater contaminants and the relative fraction of their
total mass remaining in solution (fdissolved).
FIGURE 3. Box plots showing the concentrations of caffeine,
triclosan (TCS), and triclocarban (TCC) as well as those of
microbial indicators in aquatic samples by watershed. Shown
are the median (line in box), 25th, and 75th percentile (lower
and upper hinges), a line drawn from the upper hinge to the
upper adjacent value and from the lower hinge to the lower
adjacent value (whiskers), and individual outlier values are
shown by circles. The number of observations expressed in
each box plot is displayed directly above it.
VOL. 42, NO. 9, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 3337
detected in a sample taken from Gwynns Run, were 1130
antimicrobial compounds were within the range previously
reported for a small subset of Baltimore streams (19, 20).
Raw sewage sampled at the inlet of the local wastewater
ng/L and average concentrations for triclosan and triclo-
carban of 4700 and 6100 ng/L, respectively, as determined
from long-term monitoring (24, 25). Additional statistics are
shown in the box plots presented in Figure 3 and in the
Bacteriological Analyses. All 14 streams tested positive
for each of the three bacteriological indicator organisms
examined in this study (Figure 3). Fecal coliforms, entero-
cocci, and E. coli were detected in surface water samples at
a frequency of 92, 96, and 88%, respectively; the correspond-
ing maximum concentrations of these indicators were 6.5 ×
105, 5.8 × 104, and 4.2 × 105CFU/(100 mL), respectively.
Concentrations of fecal coliforms, enterococci, and E. coli
measured in raw sewage (influent) sampled at the local
× 106CFU/(100 mL), respectively. Additional information is
presented in the Supporting Information (Table S3).
OWCs as chemical indicators of microbial risks in surface
plate counts as a benchmark. Coefficients of determination
(R2values) for triclosan and triclocarban were in the range
of 0.45-0.55, and similar results were obtained when the
inferior chemical predictor of bacterial contaminants, as
enterococci (R2) 0.55) and the weakest between caffeine
and enterococci (R2) 0.16; Figure 4). Observed trends were
consistently found regardless of the approach chosen for
treating nondetect values (Table S4).
some 13% of which are impacted by pathogens, the largest
category of contaminants together with mercury (26). The
majority of these reported microbial impairments are based
either upon levels of frank pathogens or, more commonly,
from nonpathogenic strains (26). Following release of raw
sewage into surface waters during spill events, microbial
associate with settleable particulates and leave the water
column to become part of the sediment. One report showed
microbial sorption to be particularly extensive during storm
events, when an average of 40% of fecal indicator bacteria
were contained in settleable material (27). Another study
showed a positive correlation between particle associated
bacteria and POC, and a negative correlation between free
bacteria and POC (28).
Here, the association between chemical and biological
FIGURE 4. Correlation between chemical and bacteriological markers of wastewater leakage to surface water. Chemical detected
(filled diamonds), chemical not detected (open diamonds), triclosan not detected (diamonds with crosses), and triclocarban not
detected (diamonds with “x”). Nondetect measurements are expressed as half of the limit of detection or LOD, which varied by the
3338 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 9, 2008
in greater detail by testing the hypothesis that hydrophobic
chemicals showing consistent partitioning behavior at en-
vironmental pH are preferable to nonsorptive alternatives
when monitoring for biological risks.
Caffeine is a widely used chemical indicator of microbial
risks in surface waters (17, 29–32). However, since caffeine
displays essentially no association with POC via sorption
(Figure 2), it cannot track the settleable microbial burden.
As microorganisms are removed from the water column by
gravity settling, caffeine and other nonsorptive chemical
tracers will become separated from the microbial burden of
interest. If loss by settling is extensive, as is the case during
storm events (27), the predictive value of caffeine as a
chemical marker will be diminished. False positive results
settled. In contrast, particle-active chemical markers, such
suffer from this limitation and settle to bottom sediments
along with the microorganisms.
Indeed, the monitoring data obtained for the three
watersheds under study support the selection of sorptive
OWCs over the conventionally employed marker caffeine
(Figure 4). Coefficients of determination were consistently
superior for the two hydrophobic antimicrobials, whose
concentrations in natural waters are orders of magnitude
below milligrams per liter minimum inhibitory concentra-
tions (33) that would impact the viability and survival of
were measured simultaneously in environmental samples,
and the associated R2values (derived from reported cor-
relation coefficients, R) showed a considerable range from
0.22 to unity (17, 31, 32). One report showed favorable R2
values of 1.0 and 0.97 when comparing concentrations of
caffeine to those of fecal coliforms and enterococci, respec-
tively; however, this study was limited to only four sampling
events on a single small lake (31). Another study featuring
two sampling campaigns and six sites yielded an R2value of
0.49 when correlating caffeine concentrations to the occur-
here (32). The present study investigated a large number of
streams and produced a data set of comparatively greater
geographical and temporal diversity. In addition, the simul-
taneous determination of three chemical tracers and three
unbiased comparison of caffeine to its hydrophobic tracer
2). This behavior may help to explain why triclosan overall
performed slightly less favorably than triclocarban (Figure
4). Another potential issue concerning the use of triclosan
as a chemical marker for sewage spills is the compound’s
destined for disposal into wastewater. Incorporation of
triclosan into textiles and plastics of daily use is becoming
more prevalent, which opens the possibility of nonsewage
specific detections in natural surface waters. In addition, at
least 1500 triclosan-containing personal care product for-
mulations are commercially available today (34) and some
of these may wash off of swimmers and other recreational
contamination in leaked sewage.
Triclocarban was identified as a promising chemical
marker that performed well in situ, presumably because of
its pronounced sorption behavior and insensitivity to pH in
uses as triclosan, presumably because of its limited pro-
lower solubility in water (35), and a comparatively limited
body of toxicological data (36).
for chemical monitoring of microbial risks is their environ-
prevailing in aquatic sediments. Whereas both compounds
are susceptible to aerobic biodegradation and other destruc-
tive processes (37, 38), their half-lives in anaerobic soils and
that antimicrobial agents deposited over time in sediments
may be released spontaneously during rain storms, boating,
dredging, and other activities as a result of sediment
suspension. In contrast, caffeine has the more desirable
property of disintegrating fairly quickly in ambient waters
and not accumulating to any significant extent in aquatic
with sediment suspension of bound TCC and TCS can be
minimized by monitoring these markers concurrently with
POC and other indicators of microbial risks for example
caffeine, fecal coliforms, enterococci, or E. coli.
Ideally, chemical indicators used for monitoring of
microbial risks from sewage spills to surface waters should
originate exclusively in raw wastewater, occur therein at
elevated and relatively constant levels, adsorb to microor-
during sewage treatment, be insensitive to environmental
pH changes, and attenuate in natural waters at rates
comparable to those of pathogenic microorganisms. While
the antimicrobial OWCs evaluated here do not satisfy all of
the above requirements, they certainly represent a valid
wastewater worldwide, caffeine does occur naturally in
surface waters in coffee bean producing countries, which
is another limitation. For this reason, chemical monitoring
of microbial risks may be more effective when using
hydrophobic OWCs in place of, or in conjunction with, the
traditionally employed chemical indicator, caffeine.
This research was made possible in part by the National
Institute of Environmental Health Sciences through the
Johns Hopkins University (JHU) Center in Urban Environ-
mental Health (Grant P30ES03819), and the Maryland
Cigarette Restitution Program Research Grant given to the
Johns Hopkins Medical Institutions. This publication was
developed under U.S. EPA STAR Grant No. R833002. It has
not been formally reviewed by EPA. The views expressed in
this document are solely those of the authors, and EPA does
a JHU Faculty Innovation Award for Rolf Halden, a JHU
Center for Excellence in Environmental Public Health Track-
ing Fellowship for Jochen Heidler, and a pilot project grant
from the JHU Center for a Livable Future. We thank Darin
Crew and volunteers from the Herring Run Watershed
Association for assistance in field sampling and Kathryn
Kulbicki for GIS mapping. We thank John Martin and Nick
Supporting Information Available
in the streams investigated and details of the regression
analyses in tabular form. This material is available free of
charge via the Internet at http://pubs.acs.org.
VOL. 42, NO. 9, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 3339
Literature Cited Download full-text
(1) Montemayor, M.; Valero, F.; Jofre, J.; Lucena, F. Occurrence of
Cryptosporidium spp. oocysts in raw and treated sewage and
river water in north-eastern Spain. J. Appl. Microbiol. 2005, 99,
(2) Blanch, A. R.; Belanche-Muñoz, L.; Bonjoch, X.; Ebdon, J.;
I.; Mocé, L.; Muniesa, M.; Schwartzbrod, J.; Skraber, S.; Papa-
of established and novel microbial and chemical methods for
microbial source tracking. Appl. Environ. Microbiol. 2006, 72,
(3) Chapron, C. D.; Ballester, N. A.; Fontaine, J. H.; Frades, C. N.;
Margolin, A. B. Detection of astroviruses, enteroviruses, and
adenovirus types 40 and 41 in surface waters collected and
evaluated by the information collection rule and an integrated
cell culture-nested PCR procedure. Appl. Environ. Microbiol.
2000, 66, 2520–2525.
(4) Straub, T. M.; Quinonez-Diaz, M. D.; Valdez, C. O.; Call, D. R.;
Chandler, D. P. Using DNA microarrays to detect multiple
(5) Perry, J. D.; James, A. L.; Morris, K. A.; Oliver, M.; Chilvers, K. F.;
Reed, R. H.; Gould, F. K. Evaluation of novel fluorogenic
substrates for the detection of glycosidases in Escherichia coli
and enterococci. J. Appl. Microbiol. 2006, 101, 977–985.
(6) Kolpin, D. W.; Furlong, E. T.; Meyer, M. T.; Thurman, E. M.;
Zaugg, S. D.; Barber, L. B.; Buxton, H. T. Pharmaceuticals,
streams, 1999–2000: A national reconnaissance. Environ. Sci.
Technol. 2002, 36, 1202–1211.
(7) Weigel, S.; Kuhlmann, J.; Hühnerfuss, H. Drugs and personal
care products as ubiquitous pollutants: Occurrence and dis-
tribution of clofibric acid, caffeine and DEET in the North Sea;.
Sci. Total Environ. 2002, 295, 131–141.
(8) APHA In Standard Methods for the Examination of Water &
(9) Noble, R. T.; Weisberg, S. B. A review of technologies for rapid
detection of bacteria in recreational waters. J. Water Health
2005, 3, 381–392.
(10) Stoob, K.; Singer, H. P.; Goetz, C. W.; Ruff, M.; Mueller, S. R.
Fully automated online solid phase extraction coupled directly
to liquid chromatography-tandem mass spectrometry: Quan-
tification of sulfonamide antibiotics, neutral and acidic pesti-
cides at low concentrations in surface waters. J. Chromatogr.,
A 2005, 1097, 138–147.
of a portable immunoextraction-reversed-phase liquid chro-
Chem. 2004, 76, 805–813.
S. Characterization of Escherichia coli isolates from different
fecal sources by means of classification tree analysis of fatty
(13) Hua, W.; Bennett, E. R.; Letcher, R. J. Triclosan in waste and
surface waters from the upper Detroit River by liquid chro-
matography-electrospray-tandem quadrupole mass spectro-
metry. Environ. Int. 2005, 31, 621–630.
(14) Skutlarek, D.; Exner, M.; Färber, H. Perfluorinated surfactants
in surface and drinking waters. Environ. Sci. Pollut. Res. 2006,
(15) Federighi, M.; Tholozan, J. L.; Cappelier, J. M.; Tissier, J. P.;
Jouve, J. L. Evidence of non-coccoid viable but non-culturable
count, CTC-DAPI double staining, and scanning electron
microscopy;. Food Microbiol. 1998, 15, 539–550.
(16) Buerge, I. J.; Poiger, T.; Müller, M. D.; Buser, H.-R. Caffeine, an
waters. Environ. Sci. Technol. 2003, 37, 691–700.
inputs using caffeine, indicator bacteria, and nutrients in rural
freshwater and urban marine systems. Environ. Sci. Technol.
2006, 40, 7616–7622.
(18) Halden, R. U.; Paull, D. H. Co-occurrence of triclocarban and
triclosan in U.S. water resources. Environ. Sci. Technol. 2005,
(19) Sapkota, A.; Heidler, J.; Halden, R. U. Detection of triclocarban
and two co-contaminating chlorocarbanilides in US aquatic
environments using isotope dilution liquid chromatography
tandem mass spectrometry. Environ. Res. 2007, 103, 21–29.
(20) Halden, R. U.; Paull, D. H. Analysis of triclocarban in aquatic
spectrometry. Environ. Sci. Technol. 2004, 38, 4849–4855.
(21) Enterococci in water by membrane filtration using membrane-
1600; Office of Water (4303T), U.S. Environmental Protection
Agency: Washington, D.C., 2002.
(22) Escherichia coli (E. coli) in water by membrane filtration using
membrane-thermotolerant Escherichia coli agar (mTEC); U.S.
Protection Agency: Washington, D.C., 2002.
(23) Schwarzenbach, R. P.; Gschwend, P. M.; Imboden, D. M.
Hoboken, NJ, 2003.
(24) Heidler, J.; Halden, R. U. Mass balance assessment of triclosan
removal during conventional sewage treatment. Chemosphere
2007, 66, 362–369.
(25) Heidler, J.; Sapkota, A.; Halden, R. U. Partitioning, persistence,
and accumulation in digested sludge of the topical antiseptic
2006, 40, 3634–3639.
(26) U.S. EPA. National Section 303(d) List Fact Sheet. Accessed on
Sep. 28, 2007: http://oaspub.epa.gov/waters/national_rept.
(27) Krometis, L.-A. H.; Characklis, G. W.; Simmons, O. D., III; Dilts,
M. J.; Likirdopulos, C. A.; Sobsey, M. D. Intra-storm variability
in microbial partitioning and microbial loading rates. Water
Res. 2007, 41, 506–516.
(28) Grossart, H.-P.; Brinkhoff, T.; Martens, T.; Duerselen, C.;
Liebezeit, G.; Simon, M. Tidal dynamics of dissolved and
and fall. Limnol. Oceanogr. 2004, 49, 2212–2222.
(29) Snyder, S. A.; Westerhoff, P.; Yoon, Y.; Sedlak, D. L. Pharma-
ceuticals, personal care products, and endocrine disruptors in
water: Implications for the water industry. Environ. Eng. Sci.
2003, 20, 449–469.
(30) Scott, T. M.; Rose, J. B.; Jenkins, T. M.; Farrah, S. R.; Lukasik,
J. Microbial source tracking: Current methodology and future
directions. Appl. Environ. Microbiol. 2002, 68, 5796–5803.
of human fecal coliforms at storm water outfalls. Environ. Int.
2005, 31, 1133–1140.
(32) Ferreira, A. P. Caffeine as an environmental indicator for
assessing urban aquatic ecosystems. Cad Saúde Pública 2005,
(33) Walsh, S. E.; Maillard, J.-Y.; Russell, A. D.; Catrenich, C. E.;
Charbonneau, D. L.; Bartolo, R. G. Development of bacterial
resistance to several biocides and effects on antibiotic suscep-
tibility. J. Hosp. Infect. 2003, 55, 98–107.
assess whether household use leads to bacterial resistance.
Chem. Eng. News 2005, 83, 14.
(35) Loftsson, T.; Össurardóttir, Í. B.; Thorsteinsson, T.; Duan, M.;
Másson, M. Cyclodextrin solubilization of the antibacterial
agents triclosan and triclocarban: Effect of ionization and
polymers. J. Inclusion Phenom. Macrocyclic Chem. 2005, 52,
(36) Aiello, A. E.; Larson, E. Antibacterial cleaning and hygiene
products as an emerging risk factor for antibiotic resistance in
the community. Lancet Infect. Dis. 2003, 3, 501–506.
in sewage and activated sludge. Water Res. 1975, 9, 649–654.
W. A.; McNeill, K. Aqueous photochemistry of triclosan:
and oligomerization products. Environ. Toxicol. Chem. 2005,
(39) Ying, G. G.; Yu, X. Y.; Kookana, R. S. Biological degradation of
Environ. Pollut., online early, in press; doi: 10.1016/j.envpol.
(40) Bradley, P. M.; Barber, L. B.; Kolpin, D. W.; McMahon, P. B.;
Chapelle, F. H. Biotransformation of caffeine, cotinine, and
indicators. Environ. Toxicol. Chem. 2007, 26, 1116–1121.
3340 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 9, 2008