But Long List of Safety Tests Awaits Find by Chemist

08/05/2013
Boston Globe
By Karen Weintraub

A few years ago when consumers worried that a chemical widely used in food cans and clear plastics might pose a health risk, the answer seemed obvious: Find a substitute.

Quickly, companies began mass producing bottles made with a different chemical that is believed to be safer. Canning companies, however, continued to use bisphenol-A, also known as BPA, in the thin coating that lines the inside of food cans. 

Now, a University of Massachusetts Lowell chemist has found a chemical that might work as a substitute for BPA in cans. But the long path from a chemist’s lab to a consumer’s pantry shows why BPA is still in nearly every can of beans, tomato paste, and soup — and will be for a long time.

Daniel F. Schmidt, an associate professor of plastics engineering at UMass Lowell, said he began looking for a BPA substitute out of concern for public health — particularly the health of his wife and young children. 

BPA is chemically similar to the hormone estrogen, and animal research suggests that exposure — particularly during pregnancy and early childhood — can increase the risk of breast and prostate cancers, obesity, diabetes, infertility, and behavioral problems. In one federal study, more than 92 percent of Americans were shown to have BPA in their systems, with higher concentrations in young people.

Schmidt said he wondered why bottle makers didn’t combine their substitute chemical to make can liners. The best answer he could get was that they made bottles, not can liners, and simply didn’t have the resources.

He combined the molecule, called CBDO, with epoxy and it seemed to work. The new product has most of the same characteristics as BPA-based can liners: solid at room temperature, tough, heat-resistant, doesn’t react with oil or water. And it doesn’t look like an estrogen molecule, the way BPA does, so presumably it won’t act like one, Schmidt said.

That’s a good first step, said Cheryl Watson, a professor of biochemistry and molecular biology at the University of Texas Medical Branch in Galveston who has studied BPA.

“I think it would be really great if [Schmidt] has invented something that isn’t an estrogen and works like BPA,” she said.

Last year, Watson and some colleagues drew up a five-step process that they believe is necessary for proving that a chemical is truly safe.

After looking at a computer model to ensure it is dissimilar from hormones, Watson said, Schmidt’s chemical should be put through a series of lab tests in cells, then fish or frog embryos, and finally in mice. Only then should a chemical be considered for use in food packaging.

“We’ve been fooled before,” she said.

Despite the best efforts of researchers, these chemicals keep having unexpected effects, Watson said. Sometimes chemicals that look safe at high concentrations turn out to be unsafe at low ones, she said. “That means that our testing is not up to what we know about the biology.”

Watson published a study in January showing that a BPA alternative, BPS, which is often used to coat paper receipts, is as hormonally active — and therefore as risky — as BPA.

“They didn’t test thoroughly,” Watson said about the makers of BPS. “We’re just learning how potent these things are, because our tests are more sensitive.”

Safety isn’t the only hurdle.

BPA is extremely good at what it does, and it will be tough to find another product that comes close, said John Rost, chairman of the North American Metal Packaging Alliance.

“BPA has worked better than anything in the industry for over 50 years,” Rost said.

To meet the manufacturing and filling requirements of the canning industry a coating needs to be usable in probably 1,000 different ways, he said. It needs to be able to be sprayed in some settings and rolled in others, he said, and it needs to not react to foods as varied as oil-based fish and water-based soups.

Rost said he does not believe BPA needs replacing, but would accept an alternative such as Schmidt’s, if consumers felt comfortable with it and if it also met the needs of the industry. First, it would have to be tested for at least five to seven years, Rost said.