Environmental Health Perspectives • volume 117 | number 4 | April 2009
Research | Children’s Health
Bisphenol A [BPA; 2,2-bis(4-hydroxyphenyl)
propane] is a high-production volume chemi-
cal used primarily in manufacturing polycar-
bonate plastic and epoxy resins that can be
used in impact-resistant safety equipment,
baby bottles, as protective coatings inside
metal food containers, and as composites
and sealants in dentistry (Chapin et al. 2008;
European Union 2003, 2008; Health Canada
2008; National Toxicology Program 2008).
In experimental animals, exposure to BPA
at high doses is associated with estrogen-like
effects. At doses below the putative lowest
observed adverse effect level, exposure to BPA
has resulted in increased prostate gland volume,
altered development and tissue organization of
the mammary gland, changes in mammary and
prostate gland development that may predis-
pose to neoplasia, disruption of sexual differ-
entiation in the brain, and accelerated puberty
in females (Ceccarelli et al. 2007; Della Seta
et al. 2006; Durando et al. 2007; Gioiosa
et al. 2007; Ho et al. 2006; Howdeshell et al.
1999; Laviola et al. 2005; Murray et al. 2007;
Negishi et al. 2004; Palanza et al. 2002; Ryan
and Vandenbergh 2006; Timms et al. 2005).
The interpretation of the evidence related to
the low-dose effects of BPA is a subject of
scientific debate (Chapin et al. 2008; European
Union 2003; Goodman et al. 2008; Gray et al.
2004; National Toxicology Program 2008;
vom Saal and Hughes 2005).
Widespread exposure to BPA among
the general population (Calafat et al. 2008c;
Center for the Evaluation of Risks to Human
Reproduction 2007; National Toxicology
Program 2008; Vandenberg et al. 2007) likely
results from ingesting food containing BPA
(Kang et al. 2006; Vandenberg et al. 2007).
Acute exposure to BPA may occur in occupa-
tional settings and during treatment with den-
tal materials containing BPA (Chapin et al.
2008; European Union 2003; Vandenberg
et al. 2007). BPA may also be used in the ther-
mal paper and polyvinyl chloride (PVC) indus-
tries (European Union 2003, 2008; National
Toxicology Program 2008). In turn, PVC
is used in manufacturing medical products,
including those found in neonatal intensive
care units (NICUs), such as bags containing
intravenous fluids and total parenteral nutri-
tion and tubing associated with their adminis-
tration; nasogastric and enteral feeding tubes;
respiratory masks and endotracheal tubes; and
umbilical catheters. Di(2-ethylhexyl) phtha-
late (DEHP) is a plasticizer added to PVC in
a variety of these medical products (Food and
Drug Administration 2001). Four additional
phenols, namely benzo phenone-3 (BP-3),
methyl paraben (MePB), propyl paraben
(PrPB), and triclosan (TCS), may also have
potential applications in health- and in per-
sonal care scenarios. TCS is a broad-spectrum
antimicrobicide used extensively in medical
settings (Jones et al. 2000; Weber and Rutala
2006). BP-3 is used as a sunscreen agent in
numerous cosmetics and everyday products
(Gonzalez et al. 2006). MePB and PrPB are
antimicrobial preservatives used in personal
care products and pharmaceuticals (Cashman
and Warshaw 2005). TCS, BP-3, MePB and
PrPB may be endocrine disruptors (Darbre
and Harvey 2008; Schlumpf et al. 2001; U.S.
Environmental Protection Agency 2008),
and human exposure to these compounds is
widespread in the United States (Calafat et al.
2008a, 2008b; Ye et al. 2006a). However, the
potential health effects of these compounds in
humans are largely unknown.
Except for DEHP and other phthalates,
limited data exist on the potential exposure to
environmental chemicals, including BPA and
the other four phenols, in premature infants
undergoing intensive medical treatments.
Early-life exposures are of great concern with
regard to the potential for adverse health conse-
quences throughout the life span. Because pre-
mature infants in intensive care units are both
developmentally and physiologically imma-
ture, they are a potential high-risk population
following exposure to environmental chem-
icals. We demonstrated previously that the
intensity of use of products containing DEHP
in 54 NICU premature infants is associated
with exposure to DEHP, as reflected in urinary
Address correspondence to A.M. Calafat, Division
of Laboratory Sciences, National Center for
Environmental Health, Centers for Disease Control
and Prevention, 4770 Buford Hwy., Mailstop F53,
Atlanta, GA 30341 USA. Telephone: (770) 488-
7891. Fax: (770) 488-4371. E-mail: email@example.com
We thank A. Bishop for technical assistance in the
urinary phenol measurements.
The findings and conclusions in this report are those
of the authors and do not necessarily represent the views
of the Centers for Disease Control and Prevention.
The authors declare they have no competing
Received 8 October 2008; accepted 10 December
Exposure to Bisphenol A and Other Phenols in Neonatal Intensive Care Unit
Antonia M. Calafat,1 Jennifer Weuve,2, 3 Xiaoyun Ye,1 Lily T. Jia,1 Howard Hu,4 Steven Ringer,5 Ken Huttner,6
and Russ Hauser3,7
1Centers for Disease Control and Prevention, Atlanta, Georgia, USA; 2Rush Institute for Healthy Aging, Rush University Medical Center,
Chicago, Illinois, USA; 3Department of Environmental Health, Harvard School of Public Health, Boston, Massachusetts, USA; 4Schools
of Public Health and Medicine, University of Michigan, Ann Arbor, Michigan, USA; 5Brigham and Women’s Hospital, Harvard Medical
School, Boston, Massachusetts, USA; 6Neonatology Unit, and 7Vincent Memorial Obstetrics and Gynecology Service, Massachusetts
General Hospital, Harvard Medical School, Boston, Massachusetts, USA
Objective: We previously demonstrated that exposure to polyvinyl chloride plastic medical devices
containing di(2-ethylhexyl) phthalate (DEHP) was associated with higher urinary concentrations of
several DEHP metabolites in 54 premature infants in two neonatal intensive care units than in the
general population. For 42 of these infants, we evaluated urinary concentrations of several phenols,
including bisphenol A (BPA), in association with the use of the same medical devices.
MeasureMents: We measured the urinary concentrations of free and total (free plus conjugated)
species of BPA, triclosan, benzophenone-3, methyl paraben, and propyl paraben.
results: The percentage of BPA present as its conjugated species was > 90% in more than three-
quarters of the premature infants. Intensity of use of products containing DEHP was strongly asso-
ciated with BPA total concentrations but not with any other phenol. Adjusting for institution and
sex, BPA total concentrations among infants in the group of high use of DEHP-containing products
were 8.75 times as high as among infants in the low use group (p < 0.0001). Similarly, after adjust-
ing for sex and DEHP-containing product use category, BPA total concentrations among infants in
Institution A were 16.6 times as high as those among infants in Institution B (p < 0.0001).
cOnclusiOn: BPA geometric mean urinary concentration (30.3 µg/L) among premature infants
undergoing intensive therapeutic medical interventions was one order of magnitude higher than
that among the general population. Conjugated species were the primary urinary metabolites of
BPA, suggesting that premature infants have some capacity to metabolize BPA. The differences in
exposure to BPA by intensity of use of DEHP-containing medical products highlight the need for
further studies to determine the specific source(s) of exposure to BPA.
Key wOrds: benzophenone, biomonitoring, BPA, glucuronidation, neonate, NICU, paraben,
triclosan. Environ Health Perspect 117:639–644 (2009). doi:10.1289/ehp.0800265 available via
http://dx.doi.org/ [Online 10 December 2008]
Calafat et al.
volume 117 | number 4 | April 2009 • Environmental Health Perspectives
concentrations of DEHP metabolites that were
substantially higher than those among the
general population (Green et al. 2005; Weuve
et al. 2006). In this follow-up report, we stud-
ied these premature infants’ intensity of expo-
sure to DEHP-containing medical devices in
relation to urinary concentrations of several
environmental phenols, including BPA.
Study population. In 2003, we studied a
convenience sample of 54 low-birth-weight
infants from level III NICUs (which provide
all newborn care, including mechanical and
high-frequency ventilation, surgery, and car-
diac catheterization) at two major Boston-area
(Massachusetts) hospitals (institutions A and
B), as previously described in detail (Green et al.
2005). We chose these infants to reflect a range
of diagnoses (including congenital anomalies
and developmental and metabolic abnormali-
ties) and NICU care requirements (e.g., ven-
tilation, enteral feedings, parenteral nutrition,
indwelling catheterization). The eligibility cri-
teria included a) being a patient in the NICU
at least 3 consecutive days before enrollment;
b) having a corrected gestational age (gesta-
tional age at birth plus age after birth) of ≤ 44
weeks; and c) having been born at or transferred
to either hospital between 1 March and 30
April 2003. We excluded infants with hyper-
bilirubinemia (> 20 µg/dL), which indicates
impaired hepatic enzyme function or structural
integrity (e.g., biliary atresia). We determined
by observation (we had no access to medical
records) length of stay in the NICU, primary
diagnosis, exposure to medical products, ges-
tational age, sex, and whether the infant was
breast- or formula-fed. We also collected at least
one urine sample per infant as normally dis-
carded human waste; a protocol-derived record
number linked each urine sample to an infant.
We had no access to any personal identifiable
information, and therefore did not conduct
parental interviews or seek parental consent.
The original study protocol was approved
by the Institutional Review Boards (IRB) of
Harvard School of Public Health, Brigham and
Women’s Hospital, and Massachusetts General
Hospital. In 2008, a protocol for the use of
the urine samples and data collected using the
original protocol for assessing exposure to BPA
and other phenols was deemed exempt from
IRB review at Harvard School of Public Health
according to 45 Code of Federal Regulations
46.102(d) and/or (f) (Department of Health
and Human Services 2005).
Intensity of use of products containing
DEHP. Before data collection in 2003, we
defined three levels of intensity of use of prod-
ucts containing DEHP (low, medium, and
high) based on a review of medical products
typically used in both NICUs and manufac-
turer-provided information on the products’
DEHP content. The low-DEHP category
included infants receiving primarily bottle and/
or gavage feedings. The medium-DEHP group
included infants receiving enteral feedings by
indwelling gavage tubes either continuously or
by bolus feedings; intravenous hyperalimenta-
tion by indwelling percutaneous intravenous
central catheter line or broviac or umbilical
vessel catheter; and/or continuous positive air-
way pressure by nasal prongs. The high-DEHP
category included infants receiving continu-
ous indwelling umbilical vein catheterization,
endotracheal intubation, intravenous hyper-
alimentation, and having an indwelling gavage
tube (for gastric decompression). As previously
described in detail (Green et al. 2005), one
of the study investigators observed the care
of each infant and, during this observation
period, inventoried the products used to care
for each infant. Given this inventory, a priori
we grouped infants into the categories of low-,
medium-, and high-use of DEHP-containing
products. Product use groups did not change
for any of the infants over the course of obser-
vation. As reported previously, the intensity of
DEHP-containing product use was strongly
related to the urinary concentrations of DEHP
metabolites (Green et al. 2005; Weuve et al.
2006). For example, among the infants
included in the present study, median urinary
concentrations of mono(2-ethylhexyl) phtha-
late, one of the DEHP metabolites, were 6.5,
27, and 92 µg/L, respectively, for those in the
low-, medium- and high-DEHP-containing
product use groups.
Urine collection and analysis. As described
before (Green et al. 2005), in 2003, spot
urine samples were collected at the end of
each infant’s observation period from a cot-
ton gauze placed in the infant’s diaper or from
the cotton filling of the diaper (we assumed a
100% recovery of free and conjugated urinary
species from the diaper or cotton gauze for all
compounds examined). In 2003, these urine
samples were analyzed for phthalate metabo-
lites at the Centers for Disease Control and
Prevention (Atlanta, GA) and the remaining
samples were stored under controlled condi-
tions at –40°C. In 2008, for this secondary
analysis, we measured BPA concentrations in
57 urine archived samples collected from 41
infants, for some of whom we had up to four
samples. Of the 16 replicates, four were col-
lected concurrently with the first samples (i.e.,
the diaper/cotton gauze yielded enough urine
for more than one vial), and the remaining
samples were collected between 6 and 48 hr
after the first samples. In addition, we meas-
ured BP-3, PrPB, and TCS in 59 samples
(from 42 infants), and MePB in 58 samples
(from 41 infants). At subfreezing storage
temperatures, the conjugated species of the
phenols examined are known to be stable for
6 months (Ye et al. 2007), and unpublished
data from our laboratory extends this stability
to at least 30 months. Under the controlled
sample storage conditions of this study, we
assumed that both free and total (free plus con-
jugated) species remained stable. Nonetheless,
the estimated urinary concentrations of the
free species must be interpreted with caution
because at the time of the collection of the
specimens, we did not pre-screen the sampling
materials for the presence of these phenols.
We measured the free and total urinary
concentrations of BPA, TCS, BP-3, MePB,
and PrPB using online solid-phase extraction
coupled to high-performance liquid chro-
matography (HPLC)–isotope dilution tan-
dem mass spectrometry with peak focusing
as described before (Ye et al. 2005, 2006b).
Briefly, the conjugated species of the phe-
nols in 100 µL of urine were hydrolyzed with
β-glucuronidase/sulfatase (Helix pomatia);
this deconjugation step was omitted when
measuring the concentrations of the free spe-
cies. After hydrolysis, samples were acidified
with 0.1 M formic acid; the phenols were
preconcentrated by online solid phase extrac-
tion, separated by reversed-phase HPLC, and
detected by atmospheric pressure chemical
ionization–tandem mass spectrometry. The
limits of detection (LODs)—calculated as
3S0, where S0 is the standard deviation as
the concentration approaches zero (Taylor
1987)—were 0.2 µg/L (PrPB), 0.4 µg/L
(BPA, BP-3), 1.0 µg/L (MePB), and 2.3 µg/L
(TCS). We analyzed low-concentration (~4
to ~25 µg/L) and high-concentration (~10 to
~65 µg/L) quality-control materials, prepared
with pooled human urine spiked with the
analytes of interest, with standard, reagent
blank, and infants’ samples. Samples with
concentrations of the analytes above the high-
est calibration standard were re-extracted
using less urine, and the concentrations were
calculated after adjusting for the dilution.
Assessment of other variables. Because
the urine sample collection was anonymous
and not based on review of medical records,
information on gestational age, length of stay
in the NICU, and whether the infant was
breast- or formula-fed was not available or
was incomplete for many infants. We used
the intensity of use of products containing
DEHP, which correlates with the degree of
medical support required, as a blunt surro-
gate measure for illness severity (Weuve et al.
2006). Nonetheless, because of the anony-
mous design of the study, we were unable
to distinguish levels of illness severity within
product-use group. We were able to observe
breast- and formula-feeding status of 24 of
the infants. Eleven infants were breast-fed
(or fed breast milk), of whom 3 were also
receiving formula. In contrast, 16 infants were
formula-fed (including the 3 who were also
breast-fed). We had BPA measurements for
Exposure to BPA in NICU premature infants
Environmental Health Perspectives • volume 117 | number 4 | April 2009
14 of the formula-fed infants, but only for 4
of the breast-fed infants.
Statistical analysis. Unless noted, we per-
formed the statistical analyses using the total
concentrations of the phenols. For each phe-
nol, urinary concentrations below the LOD
were assigned a value of LOD divided by the
square root of 2 (Hornung and Reed 1990).
We computed Spearman correlations between
the free and total concentrations of BPA and,
to evaluate the variability of urinary concen-
trations in specimens collected from the same
infant, between replicate measures. In addi-
tion, to explore the co-occurrence of expo-
sures to phenols and phthalates in the NICU
environment, we computed Spearman cor-
relations among urinary concentrations of the
phenols and the five phthalate metabolites we
evaluated in our previous work (Green et al.
2005; Weuve et al. 2006). For these analy-
ses we computed both crude and institution-
adjusted correlations, because institution
confounded the associations between BPA
and the DEHP metabolites. For the crude
comparisons of urinary concentrations of
each phenol by sex, institution, or DEHP-
containing product-use group, we used the
Kruskal–Wallis nonparametric test. We also
used linear mixed effects regression to com-
pare urinary concentrations of each phenol
across DEHP-containing product-use groups,
adjusting for institution and infants’ sex, and
accounting for the correlations between rep-
licate measures among the same infants. We
fit separate regression models for each phe-
nol, using the first two concentrations avail-
able for each infant, resulting in 54, 56, 56,
and 55 urine samples used in the models of
BPA, BP-3, PrPb, and MePB, respectively.
We did not fit models for TCS because too
few urine samples had detectable levels of this
phenol. To stabilize the variances of the uri-
nary phenol concentrations, we used the log-
transformed concentration as the dependent
variable in the regression models. Therefore,
when exponentiated, the estimated regres-
sion parameters in each phenol model are
interpreted as the urinary concentration for
Table 1. Distribution of the urinary concentrations of phenols (µg/L)a in hospitalized premature infants.
a given independent variable expressed as a
proportion (or multiple) of the urinary con-
centration in that variable’s reference level; for
example, in the BPA model, the exponenti-
ated regression parameter for the medium-
DEHP product-use group is the estimated
urinary BPA concentration in this group as a
proportion of that in the low-DEHP product-
use group. We conducted all analyses with
SAS software, version 9.1.3 (SAS Institute
Inc., Cary, NC).
Forty-one premature infants had at least one
urine sample available for analysis of BPA and
MePB (42 for TCS, BP-3, and PrPB). We
detected BPA in all of the first set of urine
samples collected from these infants at total
concentrations ranging from 1.6 to 946 µg/L;
the geometric mean was 30.3 µg/L (Table 1).
MePB and PrPB were also detected in all of
the samples and BP-3 in all but two samples
(Table 1). TCS was detected in approximately
19% of the samples and therefore was not
included in any further statistical analysis. Of
interest, the free species of some of these phe-
nols were detected slightly less frequently and
at substantially lower concentrations than the
conjugated species (Table 1). For example, free
BPA was detected in 92% of the samples; the
median (1.7 µg/L) and highest concentrations
(17.3 µg/L) were an order of magnitude lower
than the total concentrations. Of interest, the
concentrations of free and conjugated species
of BPA showed a linear relationship (Spearman
correlation, r = 0.86) throughout the range of
total BPA concentrations observed (Figure 1).
Because of the limited amount of urine avail-
able for analysis, the concentration of free spe-
cies could not be measured in all samples.
The infants in this study were roughly
equally distributed between the two institu-
tions, and about 38% (information on sex
was missing for one of the infants) were male
(Table 2). Fewer infants were exposed to
DEHP-containing products at low intensity
of use than at medium or high intensity of use
Two or more replicate samples were avail-
able for 14 infants (13 for BPA); we used the
first two of these samples to assess variability
in urinary concentrations within individual
infants (Table 3). The repeated urinary mea-
surements from the same infants were highly
correlated for BPA and BP-3, and also cor-
related, although to a lesser extent, for MePB
and PrPB. As expected, the correlations
between the concurrent (i.e., more than one
urine sample collected from a single diaper or
cotton gauze) measurements were excellent
(all Spearman r ≥ 0.95) (Table 3).
As has been reported (Ye et al. 2006a), we
found that institution-adjusted urinary con-
centrations of MePB and PrPB were highly
correlated (Spearman r = 0.73, p < 0.0001),
suggesting that human exposures to these para-
bens most likely share common pathways. BPA
was moderately correlated with MePB (r = 0.40,
p = 0.01). Notably, however, BPA was cor-
related with the phthalate metabolites, includ-
ing the three metabolites of DEHP among
which it was strongly correlated with DEHP
oxidative metabolites mono(2-ethyl-5-hydroxy-
hexyl phthalate (r = 0.57, p < 0.0001) and
mono(2-ethyl-5-oxohexyl) phthalate (r = 0.57,
p < 0.0002). The remaining institution-adjusted
correlations among the different phenols with
each other and with the phthalate metabolites
measured before (Green et al. 2005; Weuve
et al. 2006) were weak to moderate (Table 4).
Range NHANES 2003–2004b
No. < LOD
< LOD (0.4)
< LOD (0.4)
< LOD (0.4)
< LOD (2.3)
< LOD (2.3)
aThe total concentrations are the sum of the free plus conjugated species of each phenol. We calculated geometric means and medians if the frequency of detection was > 60%. The
estimated concentrations of free species must be interpreted with caution because at the time of the collection of the urine specimens, we did not prescreen the sampling materials
for the presence of these phenols. bData from 314 children 6–11 years of age from NHANES (National Health and Nutrition Examination Survey) 2003–2004 for BP-3 (Calafat et al. 2008a),
BPA (Calafat et al. 2008c), and TCS (Calafat et al. 2008b). cData from a group of 100 dults (Ye et al. 206).
Figure 1. Correlation between free and total BPA
urinary concentrations for 37 premature infants,
displayed in log-10 scale (r = 0.86).
Spearman r = 0.86
Urinary concentration of
total BPA (µg/L)
Urinary concentration of
free BPA (µg/L)
Calafat et al.
volume 117 | number 4 | April 2009 • Environmental Health Perspectives
Of the 41 infants who had at least one
BPA measure, gestational age data were avail-
able for only 16 of them. After dividing the
group roughly along the median gestational age
(25–27 weeks vs. 28–34 weeks), the median
total BPA concentrations was higher (Kruskal–
Wallis p-value = 0.06) for the younger
(242 µg/L) than for the older (29 µg/L) infants
(based on gestational age). For the 25 infants
for whom we had information about length of
stay in the NICU, BPA concentrations did not
appear to vary across long (14–90 days) versus
short (2–13 days) length of stay. The difference
in BPA concentrations for the breast-fed (or
fed breast milk) infants (n = 4) and the for-
mula fed infants (n = 14) was not statistically
significant (p = 0.7).
In the crude analyses, urinary BPA con-
centrations were significantly higher among
infants at institution A than at institution B
(p < 0.0001; Table 2). Concentrations of the
other phenols measured were similar across
institutions (Table 2). Urinary concentra-
tions of BPA and the other three phenols did
not vary substantially by infants’ sex, and no
consistent associations were present between
intensity of DEHP-containing product use
and urinary concentrations of any of the phe-
nols examined (Table 2).
On adjusting for infants’ sex and intensity
of use of products containing DEHP, the
association between institution and BPA con-
centrations persisted (Table 5). On average,
urinary concentrations of BPA among infants
hospitalized at institution A were almost 17
times greater than concentrations among
infants at institution B (p < 0.0001). By con-
trast, infants hospitalized at Institution A had
BP-3 urinary concentrations that were, on
average, about one-third the concentrations
among infants at institution B (p = 0.04).
Concentrations of the two parabens did not
vary consistently by DEHP-containing prod-
uct use group or institution. In these adjusted
analyses, none of the phenol concentrations
varied by infants’ sex.
The institution- and sex-adjusted asso-
ciations between the use of medical prod-
ucts containing DEHP group and urinary
concentrations were highly significant for BPA
but not for any of the other phenols (Table 5).
Compared with infants exposed at low inten-
sity to DEHP-containing products, infants
exposed at medium intensity had BPA con-
centrations that were 3.42 times as high [95%
confidence interval (CI), 1.45–8.09], and
infants exposed at high intensity had concen-
trations that were 8.75 times as high (95% CI,
3.36–22.8). The results of this sex- and institu-
tion-adjusted analysis differed from the crude
analyses, shown in Table 2, which did not
show marked differences in BPA concentra-
tion by DEHP group. Similarly, results from
linear mixed effects regression models of BPA
that did not adjust for sex or institution also
showed no significant differences in BPA con-
centration by DEHP-containing product use
group (data not shown). These different find-
ings were most likely attributable to negative
confounding by institution; infants at institu-
tion B had lower BPA concentrations but were
much more likely than infants at institution A
to be exposed at high intensity (29% vs. 7%)
to DEHP-containing products.
We frequently detected BPA and three other
phenols (BP-3, MePB, and PrPB) in this group
of premature infants. The median urinary con-
centrations of BP-3 in these infants were about
one order of magnitude lower than the concen-
trations reported by the National Health and
Nutrition Examination Survey (NHANES)
2003–2004 for children 6–11 years of age
in the U.S. general population (Calafat et al.
2008a). Similarly, the low frequency of detec-
tion of TCS among these infants suggests that,
although TCS may be in products used by
workers in health care settings, infants’ expo-
sure to TCS in NICUs does not appear to be
higher than that of the general population. We
found that the median urinary concentrations
of MePB and PrPB in the infants examined
were higher than the median observed among
a group of 100 adults in the United States (Ye
et al. 2006a), but lower than the 95th percentile
concentrations in the same group of adults;
no data exist on the extent of paraben expo-
sure among children. These results suggest that
the use of products that contain parabens in
Table 2. Median and 25th and 75th percentile concentrations of urinary phenols (µg/L) in hospitalized premature infants, by intensity of use of products containing
DEHP, institution, and sex.
DEHP-containing product use
BP-3 BPA MePB PrPB
No. 25th Median 75th p-Valuea No. 25th Median 75th p-Valuea No. 25th Median 75th p-Valuea No. 25th Median 75th p-Valuea
44.3 0.7 0.6 52 0.3 1
aKruskal–Wallis nonparametric test for differences by group.
Table 3. Spearman correlations between repeated urinary concentrations among premature infants (first
No. of infants with
≥ 2 measurements
No. of concurrent
Table 4. Institution-adjusted Spearman correlations between the total urinary concentrations of different
analytes among premature infants (all n = 41 or 42).
BP-3 –0.04 0.16 0.02 –0.10
p-Value 0.8 0.3 0.9 0.6
BPA 0.40 0.21 0.29
p-Value 0.01 0.2 0.08
MePB 0.73 0.29
p-Value < .0001 0.08
BPA MePB PrPB MEHHPa
Abbreviations: MBP, monobutyl phthalate; MBzP, monobenzyl phthalate; MEHHP, mono(2-ethyl-5-hydroxyhexyl) phtha-
late; MEHP, mono(2-ethylhexyl) phthalate; MEOHP, mono(2-ethyl-5-oxohexyl) phthalate.
aData from Green et al. (2005); Weuve et al. (2006).
Exposure to BPA in NICU premature infants
Environmental Health Perspectives • volume 117 | number 4 | April 2009
the treatment of NICU premature infants can
result in exposure levels that may be higher than
those observed among the general population.
By contrast, we found that the median uri-
nary concentrations of BPA among the infants
(28.6 µg/L) were about one order of magni-
tude higher than the median concentration
(3.7 µg/L) and almost twice the 95th percentile
concentration (16.0 µg/L) among children 6–11
years of age who were examined as part of the
NHANES 2003–2004 (Calafat et al. 2008c).
These data suggest that exposure to BPA among
the infants in our study was much higher than
among any of the populations examined in the
United States (Chapin et al. 2008; National
Toxicology Program 2008). Furthermore, that
> 90% of the BPA excreted in the urine was in
its conjugated (e.g., glucuronide, sulfate) form
excludes the possibility that the BPA concentra-
tions measured in these infants’ urine resulted
primarily from contamination and supports
the validity of our analytical data. More impor-
tant, our findings suggest that even premature
infants have some capacity to conjugate BPA,
in agreement with previous studies suggesting
that critically ill premature infants could, to a
certain extent, metabolize DEHP metabolites to
their urinary glucuronides (Calafat et al. 2004).
Furthermore, the fact that the association
between the concentrations of free and total
BPA was linear across the range of observed
total BPA concentrations suggests that satura-
tion of the enzyme(s) responsible for the conju-
gation of BPA did not occur even at total BPA
concentrations orders of magnitude higher those
reported among the general U.S. population.
Nevertheless, although shortly after birth some
hepatic enzymes involved in phase II metabo-
lism (such as UDP-glucuronosyltransferases)
can be activated even in premature infants
(Blake et al. 2005; Burchell et al. 1989;
Jansen et al. 1992; Leakey et al. 1987), these
glucuronidation pathways are not expected to
be functional at adult rates until months after
birth (de Wildt et al. 1999). Studies designed
to confirm these preliminary findings are war-
ranted to examine the percentage of free BPA
and conjugation ability of infants and young
children, including premature infants.
In the crude analyses, the median total BPA
concentrations were about one order of mag-
nitude higher among the infants 25–27 weeks
of age than among the older (28–34 weeks)
infants. Of interest, among the younger infants
(gestational age, 25–27 weeks), five were in the
medium-intensity and three were in the high-
intensity group exposed to DEHP-containing
products. By contrast, five of the older infants
(gestational age, 28–34 weeks) were in the
low-intensity and three were in the medium-
intensity group. Therefore, the observed dif-
ferences in BPA concentrations by gestational
age are likely to be attributable partially to the
treatment given to the infants. On the other
hand, the length of stay in the NICU or feeding
method (formula vs. breast milk) did not result
in significant differences in BPA concentrations.
However, because information on gestational
age, length of stay in the NICU, and feeding
type was incomplete for many infants, these
results are preliminary; the potential association
of these variables with BPA exposure should be
addressed in future research.
In this follow-up study of premature infants
receiving care in two NICUs, we found mod-
erate to strong institution-adjusted correla-
tions between urinary concentrations of BPA
and the metabolites of DEHP, indicating that
BPA and DEHP may share common exposure
sources. Consistent with this possibility, we
found strong associations between the use of
DEHP-containing medical devices and uri-
nary concentrations of BPA. Specifically, after
adjustment for infants’ sex and institution,
urinary BPA concentrations among infants
who were exposed at high intensity to DEHP-
containing products were almost nine times
as high as those in the low-intensity group.
Urinary BPA concentrations among infants
who were exposed at medium intensity to
DEHP-containing products were more than
three times as high as those in the low-intensity
group. Furthermore, urinary BPA concentra-
tions were higher among infants hospitalized at
institution A than among infants hospitalized at
institution B, regardless of DEHP product-use
group. Of interest, the urinary concentrations
of DEHP metabolites in infants of institution
B were higher than in infants of institution A
and might have reflected the extensive use of
two DEHP-containing PVC medical devices at
institution B that were used sparingly in institu-
tion A (Green et al. 2005; Weuve et al. 2006).
We have no information about the presence of
BPA in any of the medical or other products
used in the care of the infants. Therefore, we
cannot rule out that the two institutions used
products for the care of the infants that also
differed in their BPA content which may have
contributed to the observed differences.
Moreover, BPA may leach from other
baby and consumer products, including those
made from polycarbonate plastic and epoxy
resins (European Union 2008; Health Canada
2008; National Toxicology Program 2008);
unfortunately, we did not collect information
related to the use of these products. Taken
together, these findings suggest that expo-
sure to BPA in NICUs may result from the
use of specific medical products; these prod-
ucts may also be sources of exposure to other
compounds, such as DEHP. However, this
study’s small size, narrow range of descriptive
data on the infants and the products used for
their care (except for those related to DEHP
content), and lack of environmental measures
from the infants’ NICU surroundings limited
our ability to conduct more detailed analyses.
Although the data suggest the presence of
BPA in medical devices/products, we can not
determine with certainty the source(s) of BPA
because we did not test for BPA in the medi-
cal devices and products used or the dietary
liquids ingested (e.g., breast milk or formula).
Therefore, we cannot rule out that the milk or
formula the infants consumed was the source
of BPA exposure, as BPA has been detected
both in human milk and formula (National
Toxicology Program 2008). However, we have
no reason to suspect that the amounts of milk
or formula consumed by the premature infants
Table 5. Multivariable-adjusteda relative urinary concentrations of the phenolsb (95% CI) in hospitalized premature infants, by use of DEHP-containing products,
infant’s sex, and institution.
BP-3 (n = 56) BPA (n = 54)
Concentration (95% CI) p-Value Concentration (95% CI)
DEHP-containing product use
High 0.36 (0.08–1.53) 8.75 (3.36–22.8)
Medium 0.52 (0.14–1.93) 0.36c 3.42 (1.45–8.09)
Low 1 (Referent) 1 (Referent)
Male 1.61 (0.57–4.49) 0.4c 1.07 (0.55–2.08)
Female 1 (Referent) 1 (Referent)
A 0.31 (0.10–0.96) 0.04c 16.6 (7.98–34.6)
B 1 (Referent) 1 (Referent)
MePB (n = 55) PrPB (n = 56)
p-Value Concentration (95% CI) p-Value Concentration (95% CI) p-Value
aEach metabolite is represented by its own model that is adjusted for DEHP-containing-product use group, sex, and institution. bResults expressed as multiples of the concentrations in
the reference group. For example, infants in the high use of DEHP-containing products group have BPA concentrations nearly 9 times the concentrations among infants in the low use of
DEHP-containing product group. cp-Value corresponds to overall association between variable and phenol concentration.
Calafat et al.
volume 117 | number 4 | April 2009 • Environmental Health Perspectives
differed considerably by the use of medical
products containing DEHP and the institution
in which they were patients.
To our knowledge, this study is the first
to describe infants’ exposure to BPA, TCS,
BP-3, MePB, and PrPB. In conjunction with
the previous data collected on phthalates for
the same group of premature infants (Green
et al. 2005; Weuve et al. 2006), our findings
suggest that infants may be exposed during
critical periods of their development to several
potential reproductive and developmental toxi-
cants at levels higher than those reported for
the general population. However, this was an
exposure study, and we were therefore unable
to explore whether such exposures were associ-
ated with adverse health effects in these infants.
This study is also the first to demonstrate that
increasing intensity of DEHP-containing
PVC product use is proportional to exposure
to BPA, as was reflected in significant eleva-
tions in urinary concentrations of BPA and the
absence of similar elevations in urinary concen-
trations of the other four additional phenols
examined. Concerns related to BPA toxicity as
well as high BPA exposure levels in this sensi-
tive population of low-birth-weight premature
infants may justify using products that do not
contain BPA while not compromising the qual-
ity of medical care. Future research is needed to
establish the source(s) and the potential health
outcomes of these exposures to BPA.
Blake MJ, Castro L, Leeder JS, Kearns GL. 2005. Ontogeny of
drug metabolizing enzymes in the neonate. Semin Fetal
Neonatal Med 10:123–138.
Burchell B, Coughtrie M, Jackson M, Harding D, Fournel-Gigleux
S, Leakey J, et al. 1989. Development of human-liver UDP-
glucuronosyltransferases. Dev Pharmacol Ther 13:70–77.
Calafat AM, Needham LL, Silva MJ, Lambert G. 2004. Exposure
to di-(2-ethylhexyl) phthalate among premature neonates
in a neonatal intensive care unit. Pediatrics 113:e429–e434.
Calafat AM, Wong L-Y, Ye X, Reidy JA, Needham LL. 2008a.
Concentrations of the sunscreen agent benzophenone-3
in residents of the United States: National Health and
Nutrition Examination Survey 2003–2004. Environ Health
Calafat AM, Ye X, Wong LY, Reidy JA, Needham LL. 2008b.
Urinary concentrations of triclosan in the U.S. population:
2003–2004. Environ Health Perspect 116:303–307.
Calafat AM, Ye XY, Wong LY, Reidy JA, Needham LL. 2008c.
Exposure of the U.S. population to bisphenol A and 4-tertiary-
octylphenol: 2003–2004. Environ Health Perspect 116:39–44.
Cashman AL, Warshaw EM. 2005. Parabens: a review of epide-
miology, structure, allergenicity, and hormonal properties.
Ceccarelli I, Della Seta D, Fiorenzani P, Farabollini F, Aloisi AM.
2007. Estrogenic chemicals at puberty change ER alpha in
the hypothalamus of male and female rats. Neurotoxicol
Center for the Evaluation of Risks to Human Reproduction. 2007.
NTP-CERHR Expert Panel Report on the Reproductive
and Developmental Toxicity of Bisphenol A. Research
Triangle Park, NC:National Toxicology Program.
BPAFinalEPVF112607.pdf [accessed 26 November 2007].
Chapin RE, Adams J, Boekelheide K, Gray LE, Hayward SW,
Lees PSJ, et al. 2008. NTP-CERHR expert panel report
on the reproductive and developmental toxicity of
bisphenol A. Birth Defects Res B Dev Reprod Toxicol
Darbre PD, Harvey PW. 2008. Paraben esters: review of recent
studies of endocrine toxicity, absorption, esterase and
human exposure, and discussion of potential human health
risks. J Appl Toxicol 28:561–578.
Della Seta D, Minder I, Belloni V, Aloisi AM, Dessi-Fulgheri F,
Farabollini F. 2006. Pubertal exposure to estrogenic chemi- Pubertal exposure to estrogenic chemi-
cals affects behavior in juvenile and adult male rats. Horm
de Wildt SN, Kearns GL, Leeder JS, van den Anker JN. 1999.
Glucuronidation in humans—pharmacogenetic and devel-
opmental aspects. Clin Pharm 36:439–452.
Department of Health and Human Services. 2005. Protection
of Human Subjects. 45 CFR part 46. Available: http://
[accessed 23 February 2009].
Durando M, Kass L, Piva J, Sonnenschein C, Soto AM, Luque
EH, et al. 2007. Prenatal bisphenol A exposure induces
preneoplastic lesions in the mammary gland in Wistar
rats. Environ Health Perspect 115:80–86.
European Union. 2003. European Union Risk Assessment
Report: 4,4’-Isopropylidenediphenol (Bisphenol A).
European Union. Available: http://ecb.jrc.it/DOCUMENTS/
nolareport325.pdf [accessed 12 April 2008].
European Union. 2008. Updated European Risk Assessment
Report 4,4´-Isopropylidenediphenol (Bisphenol-A).
Environment Addendum of February 2008 (to be read in con-
junction with published EU RAR of BPA, 2003 for full details).
European Union. Available: http://ecb.jrc.it/documents/
phenola_add_325.pdf [accessed 20 August 2008].
Food and Drug Administration. 2001. Safety Assessment of
Di(2-ethylhexyl)phthalate (DEHP) Released from PVC
Medical Devices. Rockville, MD:Center for Devices and
Radiological Health, U.S. Food and Drug Administration.
[accessed 11 August 2003].
Gioiosa L, Fissore E, Ghirardelli G, Parmigiani S, Palanza P.
2007. Developmental exposure to low-dose estrogenic
endocrine disruptors alters sex differences in exploration
and emotional responses in mice. Horm Behav 52:307–316.
Gonzalez H, Farbrot A, Larko O, Wennberg AM. 2006.
Percutaneous absorption of the sunscreen benzophe-
none-3 after repeated whole-body applications, with and
without ultraviolet irradiation. Br J Dermatol 154:337–340.
Goodman JE, Witorsch RJ, McConnell EE, Sipes IG, Slayton TM,
Yu CJ, et al. 2008. Weight-of-evidence evaluation of repro-
ductive and developmental effects of low doses of bisphe-
nol A. Crit Rev Toxicol; doi.org/10.1080/10408440802157839
[Online 16 October 2008].
Gray GM, Cohen JT, Cunha G, Hughes C, McConnell EE,
Rhomberg L, et al. 2004. Weight of the evidence evaluation
of low-dose reproductive and developmental effects of
bisphenol A. Hum Ecol Risk Assess 10:875–921.
Green R, Hauser R, Calafat AM, Weuve J, Schettler T, Ringer
S, et al. 2005. Use of di(2-ethylhexyl) phthalate-containing
medical products and urinary levels of mono(2-ethylhexyl)
phthalate in neonatal intensive care unit infants. Environ
Health Perspect 113:1222–1225.
Health Canada. 2008. Draft Screening Assessment for Phenol,
4,4´-(1-methyl-ethylidene) bis-(Bisphenol A). Chemical
Abstracts Service Registry Number 80-05-7. Available: http://
batch2_80-05-7_en.pdf [accessed 12 September 2008].
Ho SM, Tang WY, de Frausto JB, Prins GS. 2006. Developmental
exposure to estradiol and bisphenol A increases sus-
ceptibility to prostate carcinogenesis and epigenetically
regulates phosphodiesterase type 4 variant 4. Cancer Res
Hornung RW, Reed LD. 1990. Estimation of average concentra-
tion in the presence of nondetectable values. Appl Occup
Environ Hyg 5:46–51.
Howdeshell KL, Hotchkiss AK, Thayer KA, Vandenbergh JG,
vom Saal FS. 1999. Environmental toxins: exposure to bis-
phenol A advances puberty. Nature 401:763–764.
Jansen PLM, Mulder GJ, Burchell B, Bock KW. 1992. New
developments in glucuronidation research: report of a
workshop on “Glucuronidation, Its Role in Health and
Disease.” Hepatology 15:532–544.
Jones RD, Jampani HB, Newman JL, Lee AS. 2000. Triclosan: A
review of effectiveness and safety in health care settings.
Am J Infect Control 28:184–196.
Kang JH, Kondo F, Katayama Y. 2006. Human exposure to bis-
phenol A. Toxicology 226:79–89.
Laviola G, Gioiosa L, Adriani W, Palanza P. 2005.
D-amphetamine-related reinforcing effects are reduced in
mice exposed prenatally to estrogenic endocrine disrup-
tors. Brain Research Bulletin 65:235–240.
Leakey JEA, Hume R, Burchell B. 1987. Development of multiple
activities of UDP-glucuronyltransferase in human liver.
Biochem J 243:859–861.
Murray TJ, Maffini MV, Ucci AA, Sonnenschein C, Soto AM.
2007. Induction of mammary gland ductal hyperplasias and
carcinoma in situ following fetal bisphenol A exposure.
Reprod Toxicol 23:383–390.
National Toxicology Program. 2008. NTP-CERHR Monograph
on the Potential Human Reproductive and Developmental
Effects of Bisphenol A. Available: http://cerhr.niehs.nih.
gov/chemicals/bisphenol/bisphenol.pdf [accessed 3
Negishi T, Kawasaki K, Suzaki S, Maeda H, Ishii Y, Kyuwa S,
et al. 2004. Behavioral alterations in response to fear-
provoking stimuli and tranylcypromine induced by perina-
tal exposure to bisphenol A and nonylphenol in male rats.
Environ Health Perspect 112:1159–1164.
Palanza P, Howdeshell KL, Parmigiani S, vom Saal FS. 2002.
Exposure to a low dose of bisphenol A during fetal life or
in adulthood alters maternal behavior in mice. Environ
Health Perspect 110:415–422.
Ryan BC, Vandenbergh JG. 2006. Developmental exposure to
environmental estrogens alters anxiety and spatial mem-
ory in female mice. Hormones and Behavior 50:85–93.
Schlumpf M, Cotton B, Conscience M, Haller V, Steinmann B,
Lichtensteiger W. 2001. In vitro and in vivo estrogenicity of
UV screens. Environ Health Perspect 109:239–244.
Taylor JK. 1987. Quality Assurance of Chemical Measurements.
Chelsea, MI:Lewis Publishers.
Timms BG, Howdeshell KL, Barton L, Bradley S, Richter CA,
vom Saal FS. 2005. Estrogenic chemicals in plastic and
oral contraceptives disrupt development of the fetal
mouse prostate and urethra. Proc Natl Acad Sci USA
U.S. Environmental Protection Agency. 2008. 5-Chloro-2-(2,4-
dichlorophenoxy)phenol (Triclosan): Risk Assessment for
the Reregistration Eligibility Decision (RED) Document.
Case No 2340. DP Barcode 343544. PC Code: 054901.
Washington, DC:U.S. Environmental Protection Agency.
Vandenberg LN, Hauser R, Marcus M, Olea N, Welshons WV.
2007. Human exposure to bisphenol A (BPA). Reprod
vom Saal FS, Hughes C. 2005. An extensive new literature
concerning low-dose effects of bisphenol A shows the
need for a new risk assessment. Environ Health Perspect
Weber DJ, Rutala WA. 2006. Use of germicides in the home
and the healthcare setting: is there a relationship between
germicide use and antibiotic resistance? Infect Control
Hosp Epidemiol 27:1107–1119.
Weuve J, Sanchez BN, Calafat AM, Schettler T, Green RA,
Hu H, et al. 2006. Exposure to phthalates in neonatal
intensive care unit infants: urinary concentrations of
monoesters and oxidative metabolites. Environ Health
Ye XY, Bishop AM, Reidy JA, Needham LL, Calafat AM. 2006a.
Parabens as urinary biomarkers of exposure in humans.
Environ Health Perspect 114:1843–1846.
Ye XY, Bishop AM, Reidy JA, Needham LL, Calafat AM. 2007.
Temporal stability of the conjugated species of bisphenol
A, parabens, and other environmental phenols in human
urine. J Expo Sci Environ Epidemiol 17:567–572.
Ye XY, Kuklenyik Z, Bishop AM, Needham LL, Calafat AM.
2006b. Quantification of the urinary concentrations of
parabens in humans by on-line solid phase extraction-
high performance liquid chromatography-isotope dilution
tandem mass spectrometry. J Chromatogr B Anal Technol
Biomed Life Sci 844:53–59.
Ye XY, Kuklenyik Z, Needham LL, Calafat AM. 2005. Automated
on-line column-switching HPLC-MS/MS method with peak
focusing for the determination of nine environmental phe-
nols in urine. Anal Chem 77:5407–5413.