This article was downloaded by: [Dalhousie University]

On: 15 January 2013, At: 00:10

Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer
House, 37-41 Mortimer Street, London W1T 3JH, UK

Food Additives and Contaminants

Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/tfac19

Increased migration levels of bisphenol A from

polycarbonate baby bottles after dishwashing,
boiling and brushing
a b a a
C. Brede , P. Fjeldal , I. Skjevrak & H. Herikstad
a
Næringsmiddeltilsynet for Midt-Rogaland Forusbeen 3 N-4033 Stavanger Norway
b
Norwegian Food Control Authority PO Box 8187 Dep N-0034 Oslo Norway
Version of record first published: 10 Nov 2010.

To cite this article: C. Brede , P. Fjeldal , I. Skjevrak & H. Herikstad (2003): Increased migration levels of bisphenol A
from polycarbonate baby bottles after dishwashing, boiling and brushing, Food Additives and Contaminants, 20:7, 684-689

To link to this article: http://dx.doi.org/10.1080/0265203031000119061

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditions

This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to
anyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation that the contents
will be complete or accurate or up to date. The accuracy of any instructions, formulae, and drug doses
should be independently verified with primary sources. The publisher shall not be liable for any loss, actions,
claims, proceedings, demand, or costs or damages whatsoever or howsoever caused arising directly or
indirectly in connection with or arising out of the use of this material.
 Food Additives and Contaminants, July 2003, Vol. 20, No. 7, 684–689
Increased migration levels of bisphenol A from
polycarbonate baby bottles after dishwashing, boiling and
brushing
C. Bredey*, P. Fjeldalz, I. Skjevraky and
H. Herikstady
yNæringsmiddeltilsynet for Midt-Rogaland, Forusbeen 3, N-4033
Stavanger, Norway
zNorwegian Food Control Authority, PO Box 8187 Dep, N-0034
Oslo, Norway
(Received 12 December 2002; revised 21 March 2003;
Downloaded by [Dalhousie University] at 00:10 15 January 2013
accepted 26 March 2003)
Baby bottles are often made of polycarbonate plastic.
Impurities remaining in the bottle from the monomer
bisphenol A can migrate from the plastic bottles into
baby food, thereby causing a health concern. Previous
migration testing of new baby bottles showed only trace
migration levels of the substance. In the present work,
polycarbonate baby bottles were subjected to simulated
use by dishwashing, boiling and brushing. Migration
testing performed with both new and used bottles
revealed a significant increase in migration of bisphenol
A due to use. This finding might be explained by
polymer degradation. Bisphenol A was determined in
200-ml samples of water food simulant by a method
based on solid-phase extraction followed by gas
chromatography coupled with mass spectrometry. The
detection limit was 0.1 g l1. Twelve different
polycarbonate baby bottles were tested by filling them
with hot water (100 C) for 1 h. The mean bisphenol A
level from new bottles was 0.23 þ 0.12 g l1, while
the mean levels from bottles subjected to simulated use
were 8.4 þ 4 g l1 (dishwashed 51 times) and 6.7 þ
4 g l1 (dishwashed 169 times), respectively. None
of the bottles released bisphenol A at levels that exceed
the recently established provisional tolerable daily
intake (0.01 mg kg1 body weight/day) in the
European Union.
*
To whom correspondence should be addressed.
e-mail: cato.brede@nmt-mrog.rl.no
Food Additives and Contaminants ISSN 0265–203X print/ISSN 1464–5122 online # 2003 Taylor & Francis Ltd
http://www.tandf.co.uk/journals
DOI: 10.1080/0265203031000119061
Keywords : bisphenol A, polycarbonate, baby bottles,
migration, water, solid-phase extraction, gas chroma-
tography, mass spectrometry
Introduction
Bisphenol A (BPA) is the most common monomer for
the production of polycarbonate (PC) plastic, which
has found widespread use in baby bottles. PC has
high transparency, good wear resistance and can be
sterilized in boiling water. PC baby bottles might
contain small amounts of BPA, which can migrate
to baby food and might represent a concern to health.
BPA has been shown to have oestrogenic effects in vivo
and belongs to the group of so-called xenoestrogens
(Ben-Jonathan and Steinmetz 1998). The endocrine-
disrupting effect of BPA and possible routes for
human intake were discussed in a recent WWF
report, which also contains an extensive review
of the literature (Lyons 2000). BPA was recently
re-evaluated by the Scientific Committee on Food
(SCF), resulting in the establishment of a provisional
tolerable daily intake (TDI) for BPA at 0.01 mg kg1
body weight day1 (SCF 2002). Hence, the drafted
first amendment of the consolidated EC Directive
relating to plastic materials intended to come into
contact with foodstuffs (EEC 2002) will set the
specific migration limit (SML) for BPA to 0.6 mg kg1
of food or food simulant.
One paper reported non-detectable levels of BPA
migration from PC food-contact materials when
using a detection limit of 5 mg l1 (Howe and
Borodinsky 1998), while another paper reported
detectable BPA migration below 5 mg l1 for three of
10 commercial PC food-contact articles (Kawamura
et al. 1998). Some workers have used liquid chroma-
tography coupled with sensitive detection such as
peroxyoxalate chemiluminescence (Sun et al. 2000)
 Bisphenol A from polycarbonate baby bottles 685

or electrochemical detection (D’Antuono et al. 2001), Table 1. Simulated use of bottles before migration
which resulted in the determination of BPA in food testing.
simulant samples from PC baby bottles at levels of
0.13–0.75 and 1.2 mg l1, respectively. 1.time of
testing 2.time 3.time
It is known that PC can degrade slowly by hydrolysis Type of treatment (new bottles) of testing of testing
in hot water at high pHs (Hu et al. 1998). Therefore, it
is possible that small amounts of BPA are produced No. of rounds in the 0 51 169
dishwasher machine
in the PC baby bottle materials during dishwashing
No. of rounds in 1 12 25
and sterilization in alkaline solutions or in steam. boiling water (total (1 h) (20 h) (35.75 h)
However, no evidence of such degradation was found number of hours)
in a previous study when testing BPA migration from No. of rounds with 0 13 23
PC baby bottles to infant feed and to water food brushing
simulant (Mountfort et al. 1997). The bottles were Treatment immediately boiling boiling boiling
then subjected to 20 cycles of sterilization and sub- before migration testing
sequent food use, and the applied method of analysis
had a detection limit of 30 mg l1, which is relatively
high compared with the low levels of BPA found to
migrate from new bottles by the later studies men- ing to EC Directive 85/572/EEC (1985). Each PC
Downloaded by [Dalhousie University] at 00:10 15 January 2013

tioned above. The present work developed a method
for the determination of low levels of BPA in water
food simulant. The method was based on solid-phase
extraction of 200-ml water samples followed by gas
chromatography coupled with mass spectrometry.
BPA migration from PC baby bottles was determined
before and after simulated use.
Materials and methods
Chemicals
The following solvents were of analytical grade:
methanol (VWR International AS, Oslo, Norway),
acetone (SDS, Peypin, France) and methyl-tert-butyl-
ether (Rathburn Chemicals Ltd, Walkerburn,
Scotland). The water was highly purified by a Milli-
Q gradient system (Millipore, Billerica,
Massachusetts, USA). Cleaning of glassware and
tubing was done by 99.5% puriss grade acetone
(VWR) followed by water. BPA (Fluka, Buchs,
Switzerland) was of analytical grade and used for
calibration-standard solutions. The internal standard
was 4,40-difluorobenzophenone (Fluorchem, Old
Glossop, Derbyshire, UK).
Conditions for migration testing
Water was used as a food simulant for the migration
testing, which is the proper simulant for milk accord-
baby bottle was hot-filled with 200 ml water and kept
in an oven (100 C) for 1 h. Before migration testing,
the bottles were subjected to treatment (table 1). New
bottles were rinsed by boiling water before the first
migration test. The same bottles were then subjected
to simulated use by repeated rounds in a laboratory
dishwasher at 70 C, which is slightly higher than
the temperature range of domestic dishwashers
(55–65 C). The detergent was DeconexTM, which
produced a pH of 11.5–12 in the washing water.
Hence, the pH of the wash water was kept
within the pH range of domestic detergents (pH
10.5–12). The bottles were not subjected to acid wash,
and the final wash cycle used destilled water.
Occasionally, the bottles were kept in boiling water
or brushed.
Water extraction
The 200-ml water samples were taken from the PC
bottles and transferred to empty glass flasks. A total
of 200 ml of a 0.01 mg ml1 aqueous solution of the
internal standard was added to each sample, giving a
concentration of 10 mg l1 in the water. The solid-
phase extraction (SPE) columns contained 100 mg
ENVþ, which is a highly cross-linked hydroxylated
polystyrene-divinylbenzene copolymer (IST, Mid
Glamorgan, UK). Before use, the SPE columns were
placed on a Visiprep 12 port SPE station (Supelco,
Bellefonte, Pennsylvania, USA) and conditioned with
3 ml each of acetone and methanol, followed by brief
 686 C. Brede et al.
drying by vacuum. The SPE columns were fitted with
tube adapters and Teflon tubing and then inserted
into syringe needles penetrating a rubber cork fitted
on a 5-litre vacuum flask. Ten samples were handled
simultaneously. By applying a slight vacuum, the 200-
ml water samples from the PC baby bottles were
transferred to the polymeric adsorbent at 10–
15 ml min1. After water extraction, all tubes were
dried inside with sterile cotton-tipped applicators
(Selefatrade AB, Spanga, Sweden) followed by com-
plete removal of the remaining water by applying a
flow of 99.999% nitrogen (Hydrogas and Chemicals
AS, Oslo, Norway) at 60–80 ml min1 while heating
the columns at 55 C for 25 min. Before adjustment of
the gas flow rate, three short pressure pulses were
used to remove most of the water in the columns. A
total of 2 ml acetone were applied for elution, fol-
lowed by careful evaporation close to dryness and
Downloaded by [Dalhousie University] at 00:10 15 January 2013
then solving the sample in 100 ml methyl-tert-butyl
ether.
Gas chromatography coupled with mass
spectrometry
The gas chromatograph was a Hewlett-Packard
model 5890 series II (Agilent, Palo Alto, California,
USA). A total of 1 ml of each sample extract was
injected in the split-less mode with a Hewlett-Packard
model 7673 autoinjector. The GC capillary column
was a 30-m and 0.25-mm i.d. Zebron ZB-5
(Phenomenex, Torrance, California, USA) with a
5% phenyl and 95% methylsiloxane stationary phase
of 0.25 mm thickness. The carrier gas was 99.9999%
helium (Hydrogass) pressure-programmed to keep a
constant linear velocity of 27 cm s1. Temperature
programme: 4 min at 40 C, then programmed at
10 C min1 to 300 C and held for 10 min; injector
temperature: 250 C; detector temperature: 280 C.
The mass spectrometer was a Hewlett-Packard
model 5972 Mass Selective Detector operated by
ChemStation software G1701BA v.B.01.00. Mass
spectra were recorded in the full scan mode (m/z
33–700) at a scan speed of 1.16 scan s1 by using
70 eV electron-impact ionization. For quantification
purposes, reconstructed ion chromatograms (RIC)
were made afterwards by using ions of high abun-
dance in the mass spectra for BPA (m/z 213) and 4,40-
difluorobenzophenone (m/z 123).
Results and discussion
Analytical characteristics
The method was applicable for the determination
of BPA in water food simulant with a detection limit
of 0.1 mg l1 at a signal-to-noise ratio of 3 when
recording the peak height signal from BPA-spiked
water and the noise from the baseline of neat water
samples. However, owing to the polarity of BPA, it
was essential to have an inert GC injector liner and a
non-contaminated GC column installed. Eight 200-ml
water samples spiked with 1 mg BPA gave a relative
standard deviation of 5%. A linear calibration curve
(R2 ¼ 0.997) was obtained with water samples of
the following BPA concentrations: 1, 5, 10, 20, 40,
50 and 60 mg l1.
By operating the mass spectrometer in the full-scan
modus, it was possible to use mass spectrum recogni-
tion as an additional identification parameter.
Other migration contaminants were identified in the
GC-MS chromatograms of the water extracts, which
demonstrated the potential of the method as a generic
multi-analyte method in specific migration testing
protocols.
Bisphenol A migration
Twelve different PC baby bottles from the Norwegian
market were collected for testing BPA migration to
aqueous food simulant. The new bottles all released
detectable amounts of BPA into the water. The levels
were low, in the 0.11–0.43 mg l1 range (table 2),
with an average level of 0.23  0.12 mg l1 for all
bottles. These levels correspond with previously
reported levels from new PC baby bottles exposed
to water repeatedly at 95 C for 30 min (Sun et al.
2000).
The baby bottles were then washed several times in a
dishwasher and occasionally were boiled or brushed
(table 1). This treatment was likely to simulate use for
a number of days equal to the number of rounds in
the dishwasher. It is possible that the use of a
laboratory dishwasher might exceed the wash effect
of a domestic dishwasher and thus increase the wear
and tear of the bottles. However, after more than 50
rounds in the dishwasher, no significant evidence of
 Bisphenol A from polycarbonate baby bottles 687

Table 2. Migration levels of bisphenol A from 12 polycarbonate baby bottles.

Baby bottle BPA concentration BPA concentration BPA concentration
sample at 1.time of testing at 2.time of testing at 3.time of testing
number (new bottles) (mg l1) (51 days of use) (mg l1) (169 days of use) (mg l1)

1 0.14  0.01 6.9  0.7 5.2  0.5

2 0.43  0.04 7.3  0.7 6.0  0.6
3 0.39  0.04 11.6  1.1 2.68  0.27
4 0.13  0.01 16.3  1.6 9.4  0.9
5 0.15  0.01 7.1  0.7 2.54  0.25
6 0.14  0.01 5.6  0.6 15.2  1.5
7 0.21  0.02 6.5  0.6 6.9  0.7
8 0.11  0.01 4.7  0.5 3.1  0.3
9 0.42  0.04 9.7  1.0 5.0  0.5
10 0.16  0.02 16.7  1.7 10.8  1.1
11 0.20  0.02 4.0  0.4 4.7  0.5
12 0.26  0.03 3.7  0.4 8.8  0.9
Average: 0.23  0.12 8.4  4 6.7  4
Median: 0.18 7.0 5.6
Downloaded by [Dalhousie University] at 00:10 15 January 2013

Only one sample per bottle.

Uncertainties are given as twice the standard deviation estimated from spiked water samples.

Figure 1. Concentrations of bisphenol A in water from polycarbonate baby bottles.

wear and tear was observed in the external printed
patterns or in the transparent plastic materials of the
bottles. The bottles were tested for migration of BPA
after 51 and 169 rounds in the dishwasher, which
resulted in levels of BPA in the ranges 3.7–17 and 2.5–
15 mg l1, respectively (figure 1). The average levels for
all the bottles were 8.4  4 mg l1 (51 rounds) and
6.7  4 mg l1 (169 rounds), respectively.
Double-sided t-tests were performed by comparison
of the mean BPA levels of new bottles with bottles
subjected to simulated use. The null hypotheses
(no difference in the means) were rejected at the
p ¼ 0.01 significance level when comparing the mean
level from the 1.time of testing (new bottles) with
the mean levels from both the 2.time and 3.time of
testing. Likewise, the null hypothesis was rejected
when comparing the mean level from the 1.time of
testing with the mean level from the combined
2.time and 3.time. Hence, a significant increase in
the mean BPA migration level was confirmed for
bottles subjected to simulated use compared with
new bottles.
Migration of other compounds
Other compounds migrated from PC baby bottles
into water. In general, their signals appeared higher
for new bottles compared with the same bottles after
simulated use. Typical chromatograms for one bottle
before and after simulated use are shown in figure 2.
The somewhat decreased retention time of the first
 688 C. Brede et al.
Downloaded by [Dalhousie University] at 00:10 15 January 2013
Figure 2. Total ion chromatograms of water extracts
from sample 6. (a) First time of testing (new bottles);
(b) third time of testing.
chromatogram was due to GC column cutting and
use, while a new GC column was installed before the
second chromatogram. Some of the compounds
observed in the GC-MS chromatograms were
identified by their mass spectra. For instance,
4-cumylphenol was detected in water from five of
the samples and probably originated from the PC
polymer production, where it might be used as a chain
terminator (Ash and Ash 1995). The origin of other
migrants was not further investigated.
Worst-case exposure assessment
The migration testing of the PC baby bottles at 100 C
for 1 h was considered to be more severe than real use.
Thus, it was possible to perform a worst-case exposure
assessment for the intake of BPA by babies. On aver-
age, babies consume 150 ml food kg1 body weight
day1 for the first year of life (Suri et al. 2002), which is
substantially higher than the food consumption of
adults. By substituting the EU-established TDI, pres-
ently 0.01 mg BPA kg1 body weight day1, a reduced
specific migration limit can be estimated for BPA
which is specially suited for infants up to 1 year old.
Note that at the moment there is no agreement of
using such a reduced limit in the EU. The reduced
specific migration limit for BPA, based on the average
food consumption for babies, is 67 mg BPA kg1 food.
All the PC baby bottle samples tested in the present
work exhibited migration levels lower than this
reduced migration limit. Thus, none of the samples
produced levels that exceed the TDI established by the
EU, even after extensive simulated use.
Regulatory implications
European regulations relating to plastic materials and
articles in contact with food have a special provision
for products intended for repeated use, where each
sample then should be tested three times in sequence.
A previous report on BPA migration from two new
PC baby bottles only revealed low and non-increasing
BPA migration after four successive tests (Sun et al.
2000). However, the possibility of increased BPA
migration levels by dishwashing, boiling and brushing
has now been revealed. The present work might
therefore initiate a discussion within the EU of
whether current testing practice is adequate for the
determination of the BPA migration levels of PC
articles used repeatedly in contact with foodstuffs.
Possible degradation of polymers, with a resulting
elevated level of the monomer in the food simulant,
may encourage the development of new methods for
long-term testing of materials and articles in contact
with foodstuffs.
Conclusion
A method based on solid-phase extraction followed
by gas chromatography coupled with mass spectrom-
etry has been established for determination of BPA in
water food simulant with a detection limit of
0.1 mg l1. Twelve different samples of PC baby bot-
tles were tested for migration of BPA to water. All the
 Bisphenol A from polycarbonate baby bottles
new bottles exhibited migration levels below 1 mg l1,
while bottles subjected to dishwashing, boiling
and brushing exhibited a significant increase in the
migration levels. After 51 and 169 rounds in a dish-
washer the migration levels of BPA were 3.7–17 and
2.5–15 mg l1, respectively. Although migration levels
rose on repeated use, the TDI would not be exceeded
for infants. The increased migration levels may be due
to polymer degradation.
Acknowledgements
Ingrid Nilsen and Vigdis Loen are thanked for work
at the laboratory.
Downloaded by [Dalhousie University] at 00:10 15 January 2013
References
Ash, M., and Ash, I., 1995, Plastic and Rubber Additives Electronic
Handbook (Aldershot: Gower).
Ben-Jonathan, N., and Steinmetz, R., 1998, Xenoestrogens: the
emerging story of bisphenol A. Trends in Endocrinology and
Metabolism, 9, 124–128.
D’Antuono, A., Dall’Orto, V. C., Lo Balbo, A., Sobral, S., and
Rezzano, I., 2001, Determination of bisphenol A in food-
simulating liquids using LCED with a chemically modified
electrode. Journal of Agricultural and Food Chemistry, 49,
1098–1101.
689
EEC, 1985, Commission Directive of 19 December 1985 laying
down the list of simulants to be used for testing migration of
constituents of plastic materials and articles intended to come
into contact with foodstuffs (85/572/EEC). Official Journal of
the European Communities, L372, 14–21.
EEC, 2002, Commission Directive of 6 August 2002 relating to plastic
materials and articles intended to come into contact with
foodstuffs (2002/72/EC). Official Journal of the European
Communities, L220, 18–58.
Howe, S. R., and Borodinsky, L., 1998, Potential exposure to
bisphenol A from food-contact use of polycarbonate resins.
Food additives and contaminants, 15, 370–375.
Hu, L. C., Oku, A., and Yamada, E., 1998, Alkali-catalyzed metha-
nolysis of polycarbonate. A study on recycling of bisphenol A
and dimethyl carbonate. Polymer, 39, 3841–3845.
Kawamura, Y., Koyano, Y., Takeda, Y., and Yamada, T., 1998,
Migration of bisphenol A from polycarbonate products.
Journal of the Food Hygienic Society of Japan, 39, 206–212.
Lyons, G., 2000, Bisphenol A. A Known Endocrine Disruptor. WWF
European Toxics Programme Report (Godalming, Surrey:
WWF-UK) [http://www.wwf-uk.org/filelibrary/pdf/bpa.pdf].
Mountfort, K. A., Kelly, J., Jickells, S. M., and Castle, L., 1997,
Investigations into the potential degradation of polycarbonate
baby bottles during sterilization with consequent release of
bisphenol A. Food Additives and Contaminants, 14, 737–740.
SCF, 2002, Opinion of the Scientific Committee on Food on Bisphenol A.
SCF/CS/PM/3936 Final (Brussels: European Commission.
Health & Consumer Protection Directorate-General) [http://
europa.eu.int/comm/food/fs/sc/scf/out128_en.pdf].
Sun, Y., Wada, M., AlDirbashi, O., Kuroda, N., Nakazawa, H.,
and Nakashima, K., 2000, High-performance liquid chromato-
graphy with peroxyoxalate chemiluminescence detection of
bisphenol A migrated from polycarbonate baby bottles using
4-(4,5-diphenyl-1H-imidazol-2-yl)benzoyl chloride as a label.
Journal of Chromatography B, 749, 49–56.
Suri, R., Stowe, Z. N., Hendrick, V., Hostetter, A., Widawski,
M., and Altshuler, L. L., 2002, Estimates of nursing infant
daily dose of fluoxetine through breast milk. Biological
Psychiatry, 52, 446–451.