Risk Analysis, Vol. 30, No. 3, 2010 DOI: 10.1111/j.1539-6924.2009.01345.x
Bisphenol A: How the Most Relevant Exposure Sources
Contribute to Total Consumer Exposure
Natalie von Goetz,
1,∗
Matthias Wormuth,
2
Martin Scheringer,
1
and Konrad Hungerbu¨hler
1
Bisphenol A (BPA) is an endocrine disrupting chemical that is found in human urine through-
out industrial societies around the globe. Consumer exposure pathways to BPA include pack-
aged food, household dust, air, and dental fillings. To date, information on the relative con-
tribution of the different pathways to total consumer exposure is lacking, but is key for man-
aging substance-associated risks.
We investigated the relative contributions of the pathways known to be most relevant
for nine different consumer groups. Our results suggest that the most important pathways
for infants and children are the use of polycarbonate (PC) baby bottles and for adults and
teenagers the consumption of canned food. Dental surgery can also considerably contribute
over a short time directly after the surgery.
For infants fed with PC baby bottles with mean dose rates of 0.8 µg/kg
bw
/d the highest
exposure dose rate was calculated. This dose rate is far below the tolerable daily intake of 50
µg/kg
bw
/d. However, it is of the same order of magnitude as recently reported concentrations
that caused low-dose health effects in rodents. We find a pattern of falling exposure levels
with rising age that is supported by urinary concentrations of BPA available for selected con-
sumer groups. Similarly, the exposure levels we predict are confirmed by the levels reported
in these studies.
KEY WORDS: Bisphenol A; consumer exposure; polycarbonate baby bottles; risk management; total
exposure
1. INTRODUCTION
Bisphenol A (BPA) is a broadly used industrial
chemical. With over 2 million tons produced globally
in 2003
(1)
it is considered a high production volume
chemical. It is used in many products meant for con-
sumer use: as the monomer for polycarbonate (PC) it
is present in cookware and drinking bottles, and as a
component of epoxy resins it is used in coatings for
1
Institute of Chemical and Bioengineering, ETH Zurich, Zurich,
Switzerland.
2
Harlan Laboratories Ltd., Itingen, Switzerland.
∗
Address correspondence to Natalie von Goetz, Wolfgang-Pauli-
Str. 10, 8093 Zurich, Switzerland; tel: +41 44 632 09 75; fax: +41
44 632 11 89; natalie.von.goetz@chem.ethz.ch.
food cans. Additional applications include uses for
dental surgery and medical equipment.
(2)
Studies have shown that BPA migrates from con-
sumer goods into food items.
(3−5)
Recently, BPA
has been detected in canned food,
(6)
food simulant
prepared in PC baby bottles,
(3,7)
and drinking wa-
ter.
(8)
In combination with low-dose adverse health
effects of BPA reported in mice,
(2)
in monkeys,
(9)
and in humans,
(10)
this raises concern about con-
sumer safety. Furthermore, BPA has been classified
as an endocrine disrupting chemical. Adverse effects
to the fetus are suspected to be associated with expo-
sure of pregnant women in a specific developmental
window.
(11)
In general, humans in their developmen-
tal stages (fetus, infant, child) seem to be affected
more severely than adults.
473 0272-4332/10/0100-0473$22.00/1
C©
2010 Society for Risk Analysis

474 von Goetz et al.
Several risk assessment reports have been pre-
pared by authorities in Europe,
(12)
Canada,
(13)
and
the United States
(14)
to assess the risk to consumers
emerging from regular exposure to BPA. These as-
sessments draw different conclusions regarding the
potential risk of BPA that are mainly based on the
interpretation of the toxicity data
(15)
and shall not
be discussed here. The governmental risk assess-
ments have in common that they provide an ex-
tensive overview of literature on both toxicity and
exposure data for BPA. These literature data are
combined to derive product-specific risks for var-
ious consumer groups, an approach also found in
the product-specific risk concept followed within the
REACH (Registration, Evaluation and Authoriza-
tion of CHemicals) legislation. For deriving risk mit-
igation methods, however, this important first step
of the risk assessment needs to be followed by the
second step of estimating the aggregated risk aris-
ing from the combination of various BPA sources.
Ideally, the total exposure resulting from the com-
bination of all relevant sources and pathways should
be assessed, so that the single sources can be rated
in relation to the other sources. This leads to an es-
timate of the overall risk associated with a specific
substance, where the risk contribution of the single
pathways can be tracked. This second, more specific
step of the risk assessment is lacking in all the risk as-
sessments for BPA published to date, and it will be
supported by the aggregate exposure data presented
in this study.
Furthermore, the cited risk assessment reports
evaluated the potential risk of BPA exclusively us-
ing worst-case assumptions. The value of worst-case
evaluations, however, is very much restricted to reg-
ulatory decisions on single products. Estimations of
worst-case exposure alone cannot help to decide
on suitable measures for substance management be-
cause levels derived from worst-case assumptions can
be arbitrarily high, depending on personal judgment
and on the degree of simplification in the defini-
tion of the worst case. For substance management,
which includes substance substitution as well as risk
minimization measures, mean value comparisons are
more helpful. On the one hand, substances can then
be compared to each other and, on the other hand,
the major exposure pathways for a population can be
identified and investigated in more detail.
As indicated above, an overall risk assessment
should be based on the exposure from all relevant
products and pathways, also called the total expo-
sure.
(16)
In reality, it will not be possible to cover
Table I. Assumed Exposure Fractions and Cooking Times for
Modeling Bisphenol A (BPA) Doses from Consuming
Microwave Food and Hot Drinks
Scenario f
pc
(−) f
mw
(−) f
hd
(−) t
mw
(min) t
hd
(min)
Low 0 0 0 2 1
Mean 0.2 0.05 0.2 8 5
High 1 0.5 1 20 10
all exposure pathways. Therefore, the total exposure
should be approximated by a multisource aggregated
exposure, that is, an aggregated exposure from the
most relevant products and pathways.
(17)
In this work, we present an estimation of the ag-
gregated consumer exposure to BPA that approxi-
mates the total exposure and thus can be used for the
derivation of the substance-related overall consumer
risk.
As the behavior of consumers strongly depends
on their regional background we focused our assess-
ment of the total consumer exposure to BPA on the
German and Swiss populations as an example. As the
behavior patterns of humans as well as the effects
of BPA are age dependent, we assess the mean to-
tal consumer exposure for six different age groups.
These results are compared to biomonitoring data
for German adults
(18)
and German children,
(19)
and
external exposure data for toddlers derived from a
duplicate diet study in the United States.
(20)
2. METHODS
2.1. Definition of Consumer Groups
One way to define subgroups of a human popu-
lation is to differentiate them with regard to their ex-
posure schemes. For BPA our first rough assessments
of exposure showed that the most relevant pathway
leading to consumer exposure is food intake. Conse-
quently, in the present work subgroups are defined
on the basis of age and gender, important drivers for
food intake. The subgroup of infants at the age of 0–6
months was further subdivided into breastfed infants
and infants fed with PC baby bottles in order to ac-
count for the dominant type of feeding.
The nine consumer groups defined in this study
(breastfed infants Infant I br, bottle-fed infants In-
fant I bo, Infant II, Toddler, Child, male teens Teen
m, female teens Teen f , male adults Adult m, female
adults Adult f ) are further described in Table II.