Bisphenol A: A Scientific Evaluation

Abstract

Bisphenol A (BPA) is used in the production of high-volume polycarbonate and epoxy resin compounds found in a number of consumer products, including plastic bottles and the linings of canned goods. As a result of such applications, very small amounts of BPA can migrate into food and drink. In light of reports suggesting that low doses of BPA cause estrogenic effects in laboratory animals, concerns were raised about the safety of these consumer products, particularly plastic bottles used for feeding milk to babies. To evaluate the risk, if any, from BPA, investigations were undertaken to more precisely determine human exposure levels and more carefully study the validity of the low-dose effects reported. On the basis of the most realistic studies of BPA levels in food and drink, as well as in human urine, it has been estimated that human exposures, including those of children, are very low and range from about .001 to .1 mcg/kg body weight (bw)/day. The results of the additional toxicology studies indicated that the low-dose effects could not be consistently replicated. In light of this, a number of governments and agencies brought together independent expert panels to carefully evaluate the toxicologic studies and provide regulatory guidance. These panels came to a similar conclusion, namely, that low-dose effects have not been demonstrated. They also supported the acceptable daily intake levels previously calculated on the basis of high-dose effects shown in laboratory animals. Comparing these acceptable intakes with the best exposure estimates reveals that human doses of BPA from migration of the compound into food and drink are orders of magnitude lower than acceptable daily intakes. Thus, it is very unlikely that humans, including infants and young children, are at risk from the presence of BPA in consumer products.

What Is It and How Is It Used?

Bisphenol A (BPA) is an industrial chemical that serves as a raw material for the manufacture of a variety of compounds, many of which are present in consumer goods. Its major use is as a monomer in the production of polycarbonates and epoxy resins. BPA is also used in lesser amounts in various applications, for example, as an intermediary in thermal paper production and as an antioxidant and inhibitor of end polymerization in the manufacture of poly(vinyl chloride) plastics.

Polycarbonates are used in a variety of consumer products, including a number that come into contact with food, such as returnable beverage containers, infant feeding bottles, plastic dinnerware, and plastic storage containers. In addition, they are used in nonfood consumer goods, such as eyeglass lenses, electrical equipment, household appliances, and sport safety equipment. Epoxy resins also have a number of applications that involve contact with food. In particular, these resins are part of the protective linings used in food and beverage cans and may also be used in food-processing applications, for example, as a coating applied to the interior surfaces of some vats used in wine making. Other applications of epoxy resins include epoxy resin-based paints, floorings, adhesives, protective coatings, and printed circuit boards. Similar resins are also used in some dental composites and sealants.

How Are Children and Adults Exposed to BPA?

Because BPA is used in the manufacture of a variety of high-volume consumer products, there are a number of possible routes of exposure for members of the general population. These can be divided into (1) direct and indirect environmental exposure due to the release of BPA during its production, use, and disposal; (2) exposure through leaching into food; and (3) contact with or inhalation of non-food-contact consumer products. Environmental measurements and knowledge of the properties of BPA suggest that environmental sources of BPA exposure do not contribute significantly to overall population exposure.[1]

It also does not appear that nonfood consumer products are significant sources of BPA exposure in the general public. Thus, the main concern is about exposure through food, particularly for infants who may be exposed through migration of BPA into milk from polycarbonate bottles and from migration of BPA into formula and canned infant foods from epoxy resin coatings of formula and food containers. The particular concern about infants is based on the assumptions that infants are most susceptible to any adverse effects of BPA and that infants ingest larger amounts of BPA per body weight than older children and adults.

What Is the Magnitude of BPA Exposure?

To determine whether ingestion of BPA by infants (as well as by adults and children) is a significant public health threat, the first step is to assess the amount of dietary BPA exposure. Although direct measurements of this exposure value would be ideal, they are not widely available, and so a variety of indirect approaches have been employed.[2] Exposure values may be based on:

• The results of studies of the migration of BPA into food and drink from polycarbonate bottles or food and beverage cans;

• Measurements of actual levels of BPA in canned goods; and/or

• Estimates of the amount of each type of food or drink consumed.

The results from the migration studies, measurements of BPA in food and drink, and estimates of food and drink consumption are then combined as appropriate for different age groups. The calculations are performed with assumptions that are designed to produce an exposure level that is significantly higher than would be calculated with the best science. For example, it is assumed that harsh laboratory methods of inducing migration of BPA into food and beverages reflect realistic migration values, and that the highest migration values and infant food- and drink-consumption numbers are the most appropriate to use in the calculations. Exaggerated exposure estimates are commonly used by regulators to encourage risk-management measures; they include significant margins of safety, and thus err on the side of protectiveness.

The European Commission Scientific Committee on Food took this approach in estimating human exposure to BPA in its 2002 report.[2] Its conclusion was that, with these assumptions, the daily intake of BPA by infants is about 1.6 mcg/kg bw/day. A similar approach calculated that children of 4–6 years of age consume approximately 1.2 mcg BPA/kg bw/day and adults consume about .4 mcg BPA/kg bw/day.

Another approach to evaluating exposure is to measure levels of BPA and/or its metabolic products in human urine, and to combine this with information on the metabolism of BPA in humans. Knowing how much BPA and its metabolic products are found in urine and how much BPA is converted to its metabolic products before being excreted makes it possible to calculate the amount ingested. A number of studies of urine levels in adults have been conducted,[3,4] and the results indicate that BPA intake in this population ranges from about .002 to .3 mcg/kg bw/day, with the median being much closer to the low value. The wide range of results reflects significant person-to-person variability, and close examination of the data from some of these studies also reveals that intakes vary significantly from day to day, especially in individuals having higher intakes.

A last approach, represented by just 1 study conducted by Wilson and coworkers,[5] is to make actual measurements of BPA levels in food, liquid, and the environment for a specific population, and to combine these measurements with estimates of ingestion and inhalation amounts for this population to calculate the total exposure. This was done for a group of children aged 2–5 years, and measurements were made both in the home and at the day care facility they attended. On the basis of the combination of exposures, the median intake of BPA was calculated to be about .04 mcg/kg bw/day and the value for the child with the highest exposure to BPA was .07 mcg/kg bw/day.

Taken together, the results of these different approaches to evaluating exposure reveal fairly consistent estimates of daily intakes of BPA by both children and adults. Although there is variation from individual to individual and over time, it appears that the best estimates of BPA ingestion range from about .001 to .1 mcg/kg bw/day. Reports of higher exposure estimates are based on studies that rely heavily on assumptions about consumption behavior and laboratory simulations of migration rather than direct measurements of body fluids or levels in food and drink ingested.

Although the evidence is limited, available data suggest that exposure amounts per body weight of adults and young children are similar. In addition, they indicate that the difference in BPA exposure per body weight between infants and young children is small.

What Is the Toxicity of BPA?

To determine the toxicologic implications of the above exposure values, an assessment of the doses required to produce human toxicity is mandatory. The first step in this assessment is to examine how BPA is metabolized by humans as compared with animals, so that the human relevance of the results of toxicology tests in animals can be evaluated. A number of studies of the metabolism of BPA have been performed in rodents, mainly rats.[1,6] The results of these investigations show that BPA is rapidly absorbed from the gastrointestinal tract after ingestion and is then converted to a number of metabolites, mainly BPA glucuronide, in the liver. After 2–3 days, excretion of BPA and its metabolites, mainly in the feces, is mostly complete. A very small fraction, less than 1%, is retained in the tissues. BPA can be transferred to rodent offspring through the placenta and in maternal milk, but quantities found are only a very small fraction of the amount administered to the mother.[2] There is no evidence of accumulation of BPA in the fetus.

Until recently, the results from rat metabolism studies were assumed to be directly applicable to humans. However, a recent study performed by Völkel and coworkers[7] with human volunteers indicates that humans metabolize orally administered BPA much more completely and more rapidly than rats. The results showed that BPA glucuronide is very rapidly formed and excreted in urine, and that this process is essentially complete within 24 hours. Practically none of the BPA is retained in the bodies of humans. These differences in metabolism need to be considered when extrapolating the results of toxicology studies from rodents to humans. For example, the more rapid excretion in humans leaves less time for interactions of BPA and its metabolites with tissues. For another, the very complete conversion to the BPA glucuronide means that effects in rodents due to the parent compound (BPA) are unlikely to be found in humans. The significance of such differences is explored more fully subsequently.

What Kinds of Toxic Effects Do BPA Cause?

BPA has been in use commercially for over 50 years, and workers producing this compound and its products (eg, epoxy resins) have been exposed to time-weighted average air levels to about 10 mg/m³ over decades. From this experience, it has been found that high exposures to BPA are irritating to the eye and respiratory tract, and may cause skin lesions and photosensitization of the skin. No studies reporting systemic effects were identified, and no epidemiologic studies of workers who have been exposed occupationally were found.

Thus, understanding of the toxic effects of BPA is based mainly on studies of rats and mice. Although there is general agreement about the ability of BPA to cause adverse effects in these animals when administered at high doses, its ability to cause effects at low doses has been a matter of contention.[8] This conflict has mainly centered on the possibility of adverse reproductive and developmental effects from low-dose BPA exposure, and has focused on whether or not BPA acts as a hormone disruptor or modulator at such doses.

However, before carefully examining the results of the low-dose research, it is useful to review what has been learned from the high-dose studies. In lifetime feeding studies on rats and mice conducted by the US National Toxicology Program (NTP),[9] the only sign of toxicity at high doses was reduced body weight, and this occurred only at doses greater than 50 mg/kg bw/day. Other research, including multigeneration feeding studies,[10,11] reinforce the conclusion that reduced body weight is essentially the only consistent and reproducible effect produced when high doses are administered. One study[12] showed that extremely high doses, 1250 mg/kg bw/day, led to fetal toxicity, but this finding is of limited relevance because the effects only occurred in association with significant toxicity in the mother. Even at such high doses, no fetal malformations were found.

As part of the NTP studies mentioned above,[9] a careful evaluation of possible carcinogenic effects of BPA was performed. The results indicate that BPA does not have carcinogenic potential even at very high doses. Supporting these findings are studies of the genotoxicity of BPA. Although there were some indications of genotoxicity at high doses in a limited number of studies, the weight of evidence from all of the standard assays indicates that BPA is not mutagenic. The combination of negative cancer bioassay results and negative mutagenicity assays strongly indicates that BPA is not carcinogenic in animals,[13] and provides no support for the hypothesis that this substance is carcinogenic in humans.

In recent years, however, attention has focused on the possibility that low, environmentally relevant doses of BPA can cause effects on human development and reproduction. This attention was based on general concerns about the endocrine effects of environmental contaminants and the experimental evidence that BPA binds to the estrogen receptor, although very weakly.[14] As a result, research on possible endocrine effects of BPA was initiated and resulted in reports that BPA causes (1) increases in weight and size of the prostate gland in male offspring of treated mice, (2) decreases in sperm efficiency in young mice, and (3) earlier puberty in female offspring of exposed mouse mothers.[8]

However, attempts to replicate these findings in other laboratories and as part of larger, more comprehensive investigations have been unsuccessful. No effects on prostate size or weight, on sperm quality, or on sexual development were seen in studies designed to replicate the experiments described previously.[8] In addition, no low-dose effects of any kind were found in a large-scale, 3-generation feeding study.[11] This is consistent with the lack of low-dose effects seen in another recent, large, 2-generation animal study.[10]

In light of the conflicting results, independent scientific panels have been convened to evaluate whether or not the weight of the evidence shows that BPA can cause effects on reproduction and development at low doses. One group, empanelled by the NTP in 2001, carefully examined each of the studies and came to the conclusion that “the Subpanel is not persuaded that a low dose effect of BPA has been conclusively established as a general or reproducible finding.[8]” The European Commission Scientific Committee on Toxicity, Ecotoxicity and the Environment (CSTEE) examined the same evidence and came to the conclusion that “a number of high quality studies on the reproductive and developmental effects of BPA are already available and do not support low dose effects.[15]”

Recent decisions by government agencies reinforce the conclusion that there is no consistent evidence supporting any low-dose effects. For example, the European Commission Scientific Committee on Food reached this conclusion with respect to its tolerable daily-intake value for BPA.[2]

An additional low-dose study was published recently after the evaluations described in the previous paragraphs had been completed, a study that has been heavily publicized. In this experiment conducted by Hunt and coworkers,[16] BPA was fed to female mice and the oocytes that they produced were examined. The results suggested that low doses of BPA caused abnormalities in these oocytes, which led to the hypothesis that BPA may cause effects on reproduction and development. Although attempts to replicate this particular study have not yet been reported, the lack of reproducibility of previous positive studies of BPA's reproductive and developmental toxicity potential greatly lessens the confidence that can be placed in the results of this one experiment. This confidence is further weakened by the lack of effect on reproduction and development seen in comprehensive, large-scale studies performed in recent years.[10,11] In sum, the weight of evidence as to the lack of low-dose effects expressed by a number of expert panels is unlikely to be changed by this one study.

It should also be noted that the panel reports described above did not incorporate the recent study cited earlier,[7] showing that the metabolism of BPA is much more rapid and complete in humans than in rodents. This evidence suggests that results of toxicologic studies on rodents are not directly applicable to humans because BPA and its metabolites spend less time in the bodies of humans as compared with rodents, and less of BPA and its metabolites remain in humans after exposure. Further, studies, such as those by Matthews and coworkers,[17] have shown that the glucuronide metabolite, which is formed rapidly and almost exclusively in humans, does not bind to the estrogen receptor. This suggests that BPA is even less likely to be an endocrine modulator than was concluded on the basis of the results of the rodent-metabolism studies. In addition, as pointed out by the European Commission Scientific Committee on Food,[2] published data indicate that the pregnant mouse is extremely sensitive to estrogens, as compared with humans. Thus, the metabolic and physiologic differences between rodents and humans, coupled with the toxicologic differences between BPA and BPA glucuronide, indicate that endocrine effects in humans, if they occur, are likely to require much higher exposures than those required to cause comparable effects in rodents.

What Are the Risks of BPA in Humans?

To assess the risk of BPA to humans, it is necessary to compare the doses required to cause adverse effects with the doses to which people are exposed. The ratio of these 2 values is a measure of whether toxicity is likely to occur and, if not, what the margin of safety is. When this ratio is higher, it is less likely that humans are at risk.

The absence of convincing effects of BPA at low doses suggests that the high-dose rodent studies are the most appropriate ones to use in estimating the minimum exposure that will cause toxicity in humans. The results of the 1982 NTP study mentioned previously[9] suggest that the lowest dose at which effects were observed is 50 mg/kg bw/day. This value is consistent with the results from the more recent 3-generation study by Tyl and coworkers.[11] This recent study also showed a “no-effect” level (NOEL) of 5 mg/kg bw/day. (It is important to note that these “lowest effect” levels are for reduction in body weight and are not related to reproductive or developmental toxicity.)

It is not possible to directly translate these NOELs in rodents into corresponding levels in humans because of incomplete knowledge of the biology of both types of organisms. As a result, a variety of assumptions are made in performing this extrapolation. In general, such extrapolations are performed by government agencies as part of activities aimed at reducing any human risk from likely levels of exposure. Because the aim is to be protective rather than scientifically accurate, the assumptions used are designed to overstate toxicity. Thus, the resultant values are acceptable daily intakes rather than NOELs and, as such, are lower than NOELs that may be generated with the best science.

On the basis of this approach, the US Environmental Protection Agency (EPA) has calculated its human acceptable daily-intake level, known as the Reference Dose (RfD), by dividing the rodent “lowest effect” level of 50 mg/kg bw/day by 1000. This calculation is based on the assumption that humans may be up to 10 times more sensitive than rodents to BPA exposure; that a sensitive human may be up to 10 times more sensitive than a typical human; and that significant scientific uncertainties about the effects of BPA remain due to limitations in available studies. Thus, the US EPA RfD, published in its Integrated Risk Information System, is 50 mcg/kg bw/day or .050 mg/kg bw/day.[18]

The European Commission Scientific Committee on Food[2] based its assessment on the rodent NOEL (from the 3-generation study) rather than the “lowest effect” level from the NTP study, ie, 5 mg/kg bw, and made a different assumption about the magnitude of the uncertainty factor needed to account for data gaps. As a result, it arrived at a tentative, acceptable daily-intake value, known in Europe as the Tolerable Daily Intake (TDI), of 10 mcg/kg bw/day or .010 mg/kg bw/day. It should be noted that this TDI considers effects on very young animals and is therefore applicable to young children.

As indicated previously, metabolic studies suggest that contrary to one of the fundamental assumptions that are part of the approach to calculate RfDs and TDIs, humans are likely to be less, rather than more, sensitive to BPA effects than rodents. Incorporating the results of these studies into the assessment would very likely lead to a significantly higher acceptable level of daily exposure.

As discussed in detail in the section on exposure, using the best available scientific data leads to a range of .001-.1 mcg/kg bw/day as a reasonable estimate of daily exposure. Comparing the high end of this range, .1 mcg/kg bw/day, with the RfD and TDI, it can be seen that human exposures are 100–1000 times lower than the acceptable levels calculated by government agencies. The ratio of the acceptable daily intake to the high end of the exposure-dose range is 100 for the TDI and 500 for the RfD. This value is often described as a margin of safety or a margin of exposure and in this case it is 100 or 500, depending on whether the TDI or RfD is used.

Conclusion

Although BPA is a component of a number of common consumer products, including those that come into contact with food, exposures of humans of all ages are very limited. The emergence of studies that provide more direct measures of these exposures reveals that early estimates of exposures based on more indirect approaches overstated human intake.

After thorough consideration by independent scientific panels of claims that low doses of BPA cause endocrine-related effects, such as developmental and reproductive toxicities, these panels have come to the conclusion that no such low-dose effects have been established. Although not all of the uncertainties about the toxicity of BPA have been resolved, it currently appears that the most scientifically valid assessment of the toxicity of BPA is one based on high-dose rodent studies, which show that very high exposures — up to doses that are toxic to the mother — are required to cause reproductive or developmental effects.

Government agencies have extrapolated the NOELs in these high-dose rodent studies to calculate acceptable intake levels in humans, assuming that humans are more sensitive to the effects of BPA than rodents. Comparisons of these acceptable levels with the estimated intake levels reveals that the intakes are 100–1000 times lower than acceptable intake values, leaving a large margin of safety.

Recent studies of the metabolism of BPA in humans and the toxicity of its main metabolite indicate that the margin of safety is even higher than the above analysis suggests. This research reveals that humans are even less likely than rodents to exhibit toxic effects from BPA exposure, because humans metabolize this compound more rapidly and completely than rodents. In addition, studies have shown that the main metabolite of BPA does not bind to the estrogen receptor, so that BPA metabolism reduces even further the likelihood of any endocrine effects mediated by receptor binding.

Thus, the evidence that has been collected in recent years coupled with experience from decades of occupational exposures indicates that the risk of BPA to humans of all ages is very low. It is thus very unlikely that humans, including infants, will suffer any adverse consequences, including endocrine-related effects, from current exposures to BPA in food, drink, or other consumer products.