69
Journal of Toxicology and Environmental Health, Part B, 11:69–146, 2008
Copyright © Taylor & Francis Group, LLC
ISSN: 1093-7404 print / 1521-6950 online
DOI: 10.1080/10937400701724303
DERIVATION OF A BISPHENOL A ORAL REFERENCE DOSE (RfD)
AND DRINKING-WATER EQUIVALENT CONCENTRATION
Calvin C. Willhite1, Gwendolyn L. Ball2, Clifton J. McLellan2
1Department of Toxic Substances Control, State of California, Berkeley, California, and 2NSF
International, Ann Arbor, Michigan, USA
Human exposure to bisphenol A (BPA) is due to that found in the diet, and BPA and its metabolites were detected at
parts per billion (or less) concentrations in human urine, milk, saliva, serum, plasma, ovarian follicular fluid, and
amniotic fluid. Adverse health effects in mice and rats may be induced after parenteral injection or after massive oral
doses. Controlled ingestion trials in healthy adult volunteers with 5 mg d16-BPA were unable to detect parent BPA in
plasma despite exquisitely sensitive (limit of detection = 6 nM) methods, but by 96 h 100% of the administered dose
was recovered in urine as the glucuronide. The extensive BPA glucuronidation following ingestion is not seen after
parenteral injection; only the parent BPA binds plasma proteins and estrogen receptors (ER). The hypothesis that
BPA dose-response may be described by a J- or U-shape curve was not supported by toxicogenomic data collected in
fetal rat testes and epididymes (after repeated parenteral exposure at 2–400,000 mg/kg-d), where a clear monotonic
dose-response both in the numbers of genes and magnitude of individual gene expression was evident. There is no
clear indication from available data that the BPA doses normally consumed by humans pose an increased risk for
immunologic or neurologic disease. There is no evidence that BPA poses a genotoxic or carcinogenic risk and clinical
evaluations of 205 men and women with high-performance liquid chromatography (HPLC)-verified serum or urinary
BPA conjugates showed (1) no objective signs, (2) no changes in reproductive hormones or clinical chemistry param-
eters, and (3) no alterations in the number of children or sons:daughters ratio. Results of benchmark dose (BMD10
and BMDL10) calculations and no-observed-adverse-effect level (NOAEL) inspections of all available and reproduc-
ible rodent studies with oral BPA found BMD and NOAEL values all greater than the 5 mg/kg-d NOAELs from mouse
and rat multigeneration reproduction toxicity studies. While allometric and physiologically based pharmacokinetic
(PBPK) models were constructed for interspecies scaling of BPA and its interaction with ER, multigeneration feeding
studies with BPA at doses spanning 5 orders of magnitude failed to identify signs of developmental toxicity or
adverse changes in reproductive tract tissues; the 5-mg/kg-d NOAELs identified for systemic toxicity in rats and mice
were less than the oral NOAELs for reproductive toxicity. Thus, it is the generalized systemic toxicity of ingested BPA
rather than reproductive, immunologic, neurobehavioral, or genotoxic hazard that represents the point of departure.
Using U.S. Environmental Protction Agency (EPA) uncertainty factor guidance and application of a threefold data-
base uncertainty factor (to account for the fact that the carcinogenic potential of transplacental BPA exposure has yet
to be fully defined and comprehensive neurobehavioral and immunotoxicologic evaluations of BPA by relevant
routes and at relevant doses have yet to be completed) to the administered dose NOAEL results in an oral RfD of
0.016 mg/kg-d. Assuming the 70-kg adult consumes 2 L of water each day and adopting the default 20% U.S. EPA
drinking water relative source contribution yields a 100 mg/L BPA total allowable concentration (TAC).
Popular literature, scientific investigation, and legislative activity have been stimulated by con-
cern that exposure to bisphenols, para-alkylphenols, and other xenobiotics that interact with steroid
hormone receptors contributes to infertility, impaired reproduction, precocious puberty, or
endometriosis or produces breast, vaginal, prostate, and uterine cancer (Consumer Reports, 2000;
Weitzman, 2005; Maffini et al., 2006). In 1996, the U.S. Congress passed the Food Quality
Protection Act and amended the Safe Drinking Water Act to require the U.S. Environmental Protec-
tion Agencey (EPA) to implement testing/screening strategies for endocrine-active chemicals.
Presented in part at the April 2007 U.S. EPA and U.S. Department of Defense Toxicology and Risk Assessment Conference (TRAC),
Cincinnati, OH.
The responsibility for the content of this document remains solely with NSF International. The data and conclusions reached herein
do not necessarily represent those of the State of California. Mention of trade names, proprietary products, or specific equipment does
not constitute an endorsement by NSF International or the State of California; nor does it imply that other products may not be equally
suitable.
Address correspondence to Clifton J. McLellan, NSF International, P.O. Box 130140, Ann Arbor, MI 48113–0140, USA.
E-mail: McLellan@nsf.org

70 C. C. WILLHITE ET AL.
In response, the 1998 agency’s Endocrine Disruptor Screening and Testing Advisory Committee
(EDSTAC) recommended in vitro receptor binding assays, a 3-d uterotrophic assay and a 15-d assay
in male rodents to comply with the Congressional directive. The Organization for Economic Coop-
eration and Development (OECD) (Hutchinson, 2002) and European Commission also adopted test
guidelines and tiered testing systems to detect and quantify materials like BPA with the potential to
alter endocrine function.
Bisphenol A (BPA) is one of the highest-production-volume (HPV) chemicals in the world with
total capacity >3.7 million metric tons (m.t.)/year (Anonymous, 2005). At least 400 publications
have appeared concerning health consequences associated with BPA exposure in mammals and
other animals, primarily out of concern that BPA’s chemical structure is similar to that of diethylstil-
bestrol (DES) (Roy et al., 1997; Schonfelder et al., 2002b). In vitro BPA binding with estrogen
receptors (ERα and ERβ) shows relatively low affinity (5000–26,000 times less) compared with 17β-
estradiol (E2). While the estrogenic activity of BPA was demonstrated at extraordinarily low doses
(e.g., 2–25 μg/kg), the relevance of those data to human health has been the subject of intense
debate (Ashby, 2001; Purchase, 2004; vom Saal et al., 2005a). The toxicology of BPA was previ-
ously reviewed (CERHR, 2007; Daston et al., 1997; European Commission, 2002, 2006; Goodman
et al., 2006; Gray et al., 2004; Haighton et al., 2002; Kamrin, 2004; Melnick et al., 2002; NTP,
2001; vom Saal et al., 1998, 2005a; vom Saal and Hughes, 2005; Welshons et al., 2006) and at
least two international conferences were dedicated to the substance (Chahoud et al., 2001; Richter
et al., 2007; Wetherill et al., 2007).
The U.S. EPA (1988) terminated its drinking water additives advisory program and it was
replaced by NSF International Standards 60 and 61. These standards provide protocols for evalua-
tion of products that may contribute direct (NSF/ANSI 60, 2005) or indirect (NSF/ANSI 61, 2005)
contaminants to potable water. To facilitate product evaluations under NSF/ANSI 61, the present
study was undertaken to derive a health-based concentration limit for BPA in drinking water. In
order to derive that limit, a drinking-water relative source contribution and an oral RfD must be
defined. To accomplish those objectives, BPA exposure routes and amounts were examined. Case
reports, epidemiologic data, and acute, subchronic, chronic, genotoxicity, developmental toxicity,
neurobehavioral, immunologic, and mode-of-action studies were compiled. Of the 4263 published
BPA studies found in the National Library of Medicine, only those considered most relevant to der-
ivation of an oral reference dose (RfD) and drinking-water equivalent concentration are cited here.
Using U.S. EPA test guidelines and risk assessment guidance documents to evaluate effects (1)
observed in repeated dose lab animal studies up to chronic duration, (2) observed through
repeated dosing of three generations of rats and two generations of mice, and (3) observed in other
studies relevant to human health, (4) selection of the lowest NOAELs, and (5) application of appro-
priate uncertainty factors, the drinking-water action level derived here for BPA is protective of pub-
lic health. The oral RfD and the drinking-water equivalent concentration are then compared to
health advisories and regulations promulgated by U.S., Japanese, and European agencies.
PHYSICAL AND CHEMICAL PROPERTIES
Bisphenol A has common synonyms of 2,2-bis-(p-hydroxyphenyl)-2-propane, 4,4′-isopropylidene-
2-diphenol, 4,4′-(1-methylethylidene) bisphenol, 4,4′-dihydroxydiphenyldimethylmethane, bis(4-
hydroxyphenyl)propane, dimethylmethylene-p, p′-diphenol, diphenylolpropane, and BPA. It has
the following structure (Figure 1), and physical and chemical properties (Table 1).
FIGURE 1. Bisphenol A chemical structure.