Roughly 95 percent of all baby bottles currently on the market are made of polycarbonate. As the poly in polycarbonate implies, this plastic is a polymer—a chainlike molecule constructed by linking up individual units of a common chemical. In this case, each link is a molecule of bisphenol A.

Toxicologist Koji Arizono of the Prefectural University of Kumamoto, Japan, and his colleagues tested 10 different brands of polycarbonate baby bottles—purchased in the United States, Germany, Japan, Korea, Malaysia, and the Philippines—along with other types of clear-plastic tableware. When heated, all leached bisphenol A, a chemical that mimics the hormone estrogen, into the liquids they held.

Pollutants that emulate hormones—especially estrogen—have emerged in recent years as a major environmental concern. Animal studies suggest they might increase an individual’s likelihood of developing certain cancers. During development, exposure to these environmental hormones also risks disrupting the normal growth and function of reproductive tissues and the brain. Although roughly a dozen animal or cellular studies on bisphenol A show a variety of biological effects, whether these changes have implications for people’s health remains controversial.

Last year, the Environmental Protection Agency launched a new program to begin identifying and studying such hormone-mimicking pollutants (SN: 10/17/98, p. 251).

Some of the data by Arizono’s team, which are due to be formally published in Japanese later this year, were unveiled at a press briefing in May by Thomas Natan, research director of the Washington D.C.-based National Environmental Trust. The event marked the submission of a petition to the Food and Drug Administration (FDA) asking that polycarbonate baby bottles and foodware be labeled so that concerned consumers could avoid using them.

This clear rigid plastic is a mainstay of baby bottles, picnic cutlery, unbreakable kitchen tumblers used to serve up milk and other drinks, the carboys that hold bottled water, even the cafeteria bowls used for soups.

As early as 1936, British chemists reported finding that bisphenol A functions like a weak estrogen. Because only traces of unlinked bisphenol A were present in plastics, and because the material’s early uses were not for food containers, the result was largely forgotten.

Until June 1993. That’s when some experiments being conducted by Aruna V. Krishnan and David Feldman of the Stanford University School of Medicine went awry.

The pair had been looking for signs that yeast might produce estrogenlike compounds. But every time the researchers sterilized a batch of growth medium to feed their yeast, the flasks holding the medium added something that resembled estrogen. As it turns out, the flasks were polycarbonate, and they were leaching bisphenol A.

Though the manufacturer knew its flasks could leach traces of bisphenol A upon heating, it had considered the amounts inconsequential. They were usually at or below the general limit of detection at that time—about 10 parts per billion. In the Stanford tests, however, bisphenol-A migration of just 2 to 5 parts per billion thoroughly derailed the yeast experiments (SN: 7/3/93, p. 12).

Since then, a number of labs around the world have been probing to see what else leaches bisphenol A. And the more scientists look, the more they find.

For instance, 4 years ago, the plastic resins used to line roughly half of the food cans in Europe and 85 percent of U.S. food cans proved to be a source of dietary bisphenol A.

Nicolás Olea and his colleagues at the University of Granada analyzed the contents of 20 different brands of canned foods purchased at various groceries in Europe and the United States. They included a range of popular items: corn, tomatoes, mushrooms, even artichoke hearts. Measurable quantities of bisphenol A turned up in roughly half of all foods sampled (SN: 6/3/95, p. 341).

What’s troubling, Olea noted, is that consumers have no way of knowing whether their canned goods have been packed in a plastic-lined container. That information was not provided on the label.

One year later, Olea’s team reported finding that bisphenol A leaches from some of the plastic resins commonly used to seal tooth surfaces. While these sealants reduce the risk that treated teeth will develop cavities, the Spanish researchers pointed out that newly applied sealants leave measurable quantities of bisphenol A in the mouth. It turned up in samples of saliva collected an hour after treatment—and in one case, 2 years later (SN: 4/6/96, p. 214).

Food and Drug Administration chemists were next to weigh in on bisphenol A. Two years ago, they reported that some unbound bisphenol A leached from polycarbonate baby bottles and juice cups that had been heated. There findings also suggested that some of the heating had actually started degrading the plastic—freeing up additional bisphenol A for leaching, said John E. Biles, one of the study’s authors (SN: 10/18/97, p. 255).

However, the FDA data indicated that the amount of bisphenol A that leached into a fatty material—designed to simulate infant formula—was quite small. Now, Arizono’s group has employed a sensitive new analytical technique to assay for the hormone mimic. Their data indicate that through normal use, polycarbonates leach more bisphenol A than reports by others had indicated. The concentrations of bisphenol A that result are in the same range as those that caused abnormalities in rats.

Most earlier studies investigated bisphenol-A leaching by new laboratory ware, new dental sealants, or new bottles. To probe how wear and tear might affect the migration of this chemical, Arizono’s team compared rates of bisphenol A release from shiny new baby bottles and from bottles that were hazy and crackled, reflecting years of use.

Because heating, such as the boiling used to sterilize bottles, facilitates bisphenol-A release, Arizono put 90 °C water in contact with the polycarbonate materials for 30 minutes and then measured how much bisphenol A ended up in the water.

For new baby bottles, the water picked up between 1 and 3.5 parts per billion (ppb) bisphenol A. Water heated in used but relatively clear bottles sometimes picked up as much as 6.5 ppb. Water in very worn and heavily scratched bottles acquired between 10 and 28 ppb of the compound.

Coming from a culture that esteems its soup, Arizono then filled polycarbonate soup bowls with 75 °C water. After 30 minutes, this water contained up to 2 ppb bisphenol A.

Why use water for all these studies when one earlier FDA study went so far as to assay migration from baby bottles into juice and infant formula? "Unfortunately," says Natan, such foods can "hide BPA from detection," masking the magnitude of any migration. He says Arizono stayed with water to be sure he could detect even small traces of the adulterant.

Natan’s assessment was "reasonable," an FDA scientist acknowledged, speaking on the condition that he not be named. When testing "a more complex matrix [than water]—the apple juice, or infant formula—we have a much higher detection limit," he explained. In fact, he pointed out, when working with those foods, FDA chemists had difficulty finding adulterant concentrations any lower than 100 ppb. However, he told Science News OnLine, foods or simulated foods "generally exaggerate what will actually migrate into a real food"—meaning it should identify a worst-case upper bound on risk, provided any migration exceeds the detection threshold.

Since those earlier studies, FDA’s detection threshold has been dropping. In one analysis of infant formula, published a year ago, it found 13 to 15 parts per billion of bisphenol A in the drink, with most of the hormone mimic coming from the cans’ internal coating. Newer, unpublished data by the agency’s chemists have also detected the migration of bisphenol A into some vegetables from food-can linings. Here, the foods accumulated between 5 and 39 ppb of the pollutant, the FDA reports.

The new Japanese study also looked at bisphenol-A migration from the resin used to line food cans. In Japan, vending machines dispense tea and coffee. Arizono explains that these canned drinks have become the Japanese snack-drink corollary to Coke and Pepsi in the United States. Though soft drinks stored in plastic-lined cans picked up less than 1 ppb of bisphenol A, oolong tea acquired at least 7 ppb—and coffee a whopping 90 to 127 ppb. The manufacturer reformulated its coffee containers as a result of this study, Arizono notes.

Concerned by these new data, together with bisphenol-A toxicity studies conducted in animals, the National Environmental Trust joined with 11 other consumer, health, religious and environmental organizations in petitioning FDA to:

• promptly identify all constituents of plastic food containers having the potential to migrate into foods to which children are routinely exposed,
• implement a strategy to eliminate or reduce children’s exposure to such materials, especially bisphenol A,
• allow leached materials from plastic to remain in the food supply "only after [manufacturers] have provided substantial affirmative evidence of safety—a burden that is not to be met upon a showing merely of an absence of evidence of harm,"
• and work with other federal agencies, such as the EPA and the National Toxicology Program, to investigate possible low-dose effects of such materials leaching from plastics.

To date, FDA has not formally responded to the petition.

The Society of the Plastics Industry (SPI), however, has responded. It challenges the consortium’s campaign, saying: "Over four decades of research show that polycarbonate food containers and baby bottles are safe." Moreover, one spokesman notes, since parents have made the switch from glass to plastic drink bottles, children suffer fewer dangerous accidents. Finally, SPI argues that small quantities of bisphenol A pose no health hazard, so the leaching of trace amounts should not be used to scare consumers.

In fact, there is considerable difference of opinion about the significance of biological changes seen in animals exposed to bisphenol A. In some cases, adults that had been exposed to the hormonelike substance during fetal development showed curious abnormalities. For instance, Frederick S. vom Saal at the University of Missouri-Columbia has reported that the prostates of some mice fetally exposed to 2 ppb bisphenol A were larger than normal, though no cancers were present. So, arguments have erupted on the significance of such changes.

Despite several attempts, plastics industry scientists have also failed to replicate vom Saal’s findings. There has been considerable acrimony at scientific meetings between industry scientists and vom Saal about whether they are all using the appropriate methodology to probe for effects.

Clearly, the jury is still out.

Until a verdict is returned, what can a cautious parent do?

The National Environmental Trust recommends switching to baby bottles and foodware that is not made from polycarbonates. For instance, baby bottles manufactured from a pliable, milky-colored plastic contain no polycarbonates.

Natan advocates that people who prefer to go on using polycarbonates should consider discarding any that show obvious signs of wear, such as a somewhat cloudy, crackled appearance. He also recommends heating foods and drinks outside of the plastics and then transferring them into the plastic only after they are cool enough to eat or drink.