Summary of the toxicology of bisphenol A ................................................ 115
Possible pathways of bisphenol A exposures in the general population.................................. 116
Available online at www.sciencedirect.comBiomonitoring in exposure assessment .................................................. 116
Biotransformation and toxicokinetics of bisphenol A ........................................... 117
Analytical methods for analysis of bisphenol A and bisphenol A metabolites in biological media .................... 119
Enzyme-linked immunosorbent assays (ELISAs) ........................................... 122
Gas chromatography/mass spectrometry (GC/MS) .......................................... 122
HPLC with detectors other than mass spectrometry.......................................... 123
HPLC-FLD after fluorophore derivatization.............................................. 124
Liquid chromatography/mass spectrometry (LC/MS) ......................................... 124
Summary of analytical methodology ................................................. 125
Reported concentrations of bisphenol A in human blood in non-intentionally exposed individuals.................... 126
Reported concentrations of bisphenol A in human urine after controlled exposure and in the general population ............ 126
Discussion and conclusions ........................................................ 130
Acknowledgments ............................................................. 131A due to rapid biotransformation and the very rapid excretion result in severe limitations in the use of reported blood levels of bisphenol A for exposure
assessment. Due to the rapid and complete excretion of orally administered bisphenol A, urine samples are considered as the appropriate body fluid for
bisphenol A exposure assessment. In urine samples from several cohorts, bisphenol A (as glucuronide) was present in average concentrations in the range
of 1–3
g/L suggesting that daily human exposure to bisphenol A is below 6
g per person (b0.1
g/kg bw/day) for the majority of the population.
' 2007 Elsevier Inc. All rights reserved.
Keywords: Bisphenol A; Human exposure; Biomonitoring; Exposure assessment; Risk assessment
Contents
Introduction ................................................................ 115Review
Human exposure to bisphenol A by biomonitoring: Methods, results and
assessment of environmental exposures
Wolfgang Dekant
a,


, Wolfgang Völkel
b
a
Department of Toxicology, University of Würzburg, Germany
b
Bavarian Health and Food Safety Authority, Environmental Medicine/Biomonitoring, Munich, Germany
Received 5 September 2007; revised 27 November 2007; accepted 2 December 2007
Available online 14 December 2007
Abstract
Human exposure to bisphenol A is controversially discussed. This review critically assesses methods for biomonitoring of bisphenol A exposures and
reported concentrations of bisphenol A in blood and urine of non-occupationally (“environmentally”) exposed humans. From themanymethods published
to assess bisphenol A concentrations in biological media, mass spectrometry-based methods are considered most appropriate due to high sensitivity,
selectivity and precision. In human blood, based on the known toxicokinetics of bisphenol A in humans, the expected very low concentrations of bisphenol
Toxicology and Applied Pharmacology 228 (2008) 114–134
www.elsevier.com/locate/ytaapReferences ................................................................. 131


Corresponding author. Department of Toxicology, University of Würzburg, Versbacher Strasse 9, 97078 Würzburg, Germany. Fax: +49 931 20148865.
E-mail address: dekant@toxi.uni-wuerzburg.de (W. Dekant).
0041-008X/$ - see front matter ' 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.taap.2007.12.008

pplIntroduction
Bisphenol A is primarily used to make polycarbonate plastic
and epoxy resins, which are widely used for a variety of appli-
cations such as polycarbonate baby feeding bottles and epoxy
food-can linings. Although of lesser importance, other uses have
been reported, for example as an antioxidant in polyvinylchlor-
ide production (EU, 2003). Due to the use of bisphenol A to
manufacture products used in many applications, it has been
speculated that human exposures to bisphenol A may be wide-
spread and it has been postulated that these exposures may reach
high levels (vom Saal and Hughes, 2005; vom Saal et al., 2005;
Vandenberg et al., 2007). Bisphenol A is a weak estrogen and has
been implicated to cause a variety of effects on reproduction and
development in animals, including effects at doses well below
those showing adverse effects in routine toxicity studies (vom
Saal et al., 2005). Moreover, bisphenol A blood levels deter-
mined in humans were statistically associated with human dis-
eases such as ovarian dysfunction or endometrial hyperplasia
and recurrent miscarriage (Sugiura-Ogasawara et al., 2005;
Takeuchi et al., 2004a; Berkowitz, 2006; Politch, 2006; Hiroi
et al., 2004). Therefore, some publications have called for an
updated risk assessment (vom Saal and Hughes, 2005; vom Saal
et al., 2005).
In risk assessment, exposure and hazard (potential toxicity)
are separately assessed. Integration of hazard assessment and
exposure with dose–response for toxicity is then translated into
risk assessment (Henry, 2003; MacDonald, 2004). The potential
toxicity of bisphenol A has been intensively addressed with
studies performed along testing guidelines and following good
laboratory practices (GLP) available to cover all of the major
toxicity endpoints relevant for hazard assessment (EU, 2003;
Goodman et al., 2006; Gray et al., 2004; Willhite et al., 2008).
The exposure of the general population to bisphenol A has been
assessed in a number of studies or assessments by regulatory
agencies or in publications (EU, 2003; Kang et al., 2006;
Miyamoto and Kotake, 2006; EFSA, 2006; Fujimaki et al.,
2004; SCF, 2002; Thomson and Grounds, 2005; Wilson et al.,
2007). However, many of the assessments have used data on
bisphenol A concentrations in environmental media and in food
items, or data on the migration of bisphenol A into food, as a
basis for exposure assessment. In combination with food con-
sumption data, this information served as a basis for indirectly
estimating human exposures to bisphenol A. Due to the
development of highly sensitive analytical techniques, the use
of “biomonitoring” is playing a more and more important role in
exposure assessment since, when incorporating knowledge on
toxicokinetics and biotransformation of the compound of
interest and the limitations of analytical methods, biomonitoring
gives more exact information on actual exposure to an agent as
compared to indirect assessments. Large programs to address
population exposures to environmental chemicals using biomo-
nitoring have been started (Angerer et al., 2007; Boogaard,
2007; Calafat et al., 2006; Needham et al., 2007; Pirkle et al.,
W. Dekant, W. Völkel / Toxicology and A2005; Yang et al., 2006b).
This review will critically assess the suitability of analytical
methodology available for the quantitative determination ofbisphenol A in the low concentrations expected in urine and
blood samples from non-occupationally exposed humans and
the available information on concentrations of bisphenol A in
blood and urine from such populations. This information is then
translated into an assessment of human exposure to bisphenol A
from environmental sources.
Summary of the toxicology of bisphenol A
Due to the focused scientific and public attention to
chemicals that may mimic endogenous hormone action and
thus interfere with endocrine organ function, many investiga-
tions assessing the potential toxicity of bisphenol A have been
published in recent years. Due to the high production volumes,
the toxicity of bisphenol A has been intensively studied since
the 1970s. Bisphenol A is not genotoxic, not carcinogenic and
the guideline conforming repeated dose toxicity studies,
including studies on reproductive and developmental toxicity
covering a wide dose range, showed adverse effects only at
dosesN50 mg/kg bw/day (summarized in: EU, 2003; Goodman
et al., 2006; Gray et al., 2004). A number of non-guideline and
non-GLP studies, often using low numbers of animals per
experimental group, have reported effects of bisphenol A
administration on a variety of parameters regarding reproductive
and developmental toxicity endpoints inmuch lower dose ranges
(2 to 20
g/kg bw/day), which were claimed to be due to the
weak estrogenicity of bisphenol A. The dose ranges in which
these effects have been reported have also been claimed to be
“environmentally relevant” (Sheehan, 2000; vom Saal et al.,
2005) or to result in blood levels of “unconjugated” bisphenol A
close to that encountered in humans after environmental
exposure (Keri et al., 2007).
The relevance of the low dose effects of bisphenol A in
animals for human hazard assessment is unclear, since the other
toxicity studies, often using more powerful study designs with
larger numbers of animals per dose group, a larger number of
dose levels and stringent quality control, were unable to repeat
the findings of low-dose effects after bisphenol A exposures
(Ashby et al., 1999; Cagen et al., 1999; Ema et al., 2001; Tyl
et al., 2002). Moreover, the weak estrogenicity of bisphenol A,
as assessed by special protocols, has been known for decades
(Bitman and Cecil, 1970; Dodds and Lawson, 1936) and
humans are exposed to a variety of natural chemicals with weak
estrogenicity (Bolt et al., 2001; Safe, 2004, 2000). Therefore,
the studies reporting low-dose effects of bisphenol A were not
considered sufficiently robust to serve as pivotal points in risk
assessment (EFSA, 2006). To derive a tolerable daily intake
(TDI) for bisphenol A, the European Food Safety Authority
(EFSA) mainly relied on two rodent multi-generation studies
with dose levels covering a wide dose-range including low dose
exposures, and on species differences in toxicokinetics (see
below). In a three-generation rat study, significant reductions in
adult body weight and in pup body and organ weights were
observed at a dose of 50 mg bisphenol A/kg bw/day and above
115ied Pharmacology 228 (2008) 114–134(Tyl et al., 2002). In a two-generation study in the mouse, an
increased incidence of hepatocyte hypertrophy of mild to
minimal severity after doses of 50 mg bisphenol A/kg bw/day