NEW YORK - A study out Monday finds no clear evidence that people's exposure to bisphenol A, a controversial chemical in plastics, is related to their risk of diabetes.
In a study of 3,400 Chinese adults age 40 and up, researchers found no clear-cut connection between people's urinary levels of bisphenol A - better known as BPA - and their odds of having diabetes.
The one-quarter with the highest BPA levels did have a higher risk of diabetes than the quarter with the lowest levels. But the researchers failed to find a "dose-response" relationship - where the risk of diabetes would steadily climb as BPA exposure did.
And the researchers say that means they cannot conclude there is a BPA-diabetes connection.
But the findings, published in the Annals of Internal Medicine, are far from settling the issue of BPA's possible health effects. And researchers not involved in the study questioned the authors' conclusions.
BPA is a so-called endocrine disruptor, which means it may affect normal hormone activity in the body.
It's also ubiquitous. BPA has been used for decades to make hard plastic containers, and linings for metal food and drink cans. Research suggests that most people, including about 95 per cent of Americans, have some amount of BPA in their blood.
Recent animal studies have hinted that the chemical could play a role in certain cancers, heart disease and abnormal brain development in children. Due to the controversy over BPA, the major manufacturers of infant bottles and feeding cups in the US have stopped using the chemical.
But BPA's true effects in humans remain unknown. Two large studies have found a link between higher BPA levels and higher heart disease risk. And a 2008 study found that of Americans in a government health survey, those with higher BPA levels showed a higher diabetes risk.
None of that proves cause-and-effect, however.
And these latest findings "do not confirm" a BPA-diabetes connection, write the researchers, led by Dr Guang Ning of Shanghai Jiao-Tong University School of Medicine.
Of the 3,423 men and women studied, 32 per cent had type 2 diabetes. When the researchers separated all participants into four groups based on urinary BPA level, they found that the group with the highest levels was 37 per cent more likely to have diabetes than the group with the lowest BPA concentrations.
That was with other factors - like age, weight and kidney function - taken into account.
On the other hand, Ning's team found, the study group with the second-highest BPA levels showed no increase in their diabetes risk, while the one with the second-lowest did.
So there was no "clear, monotonic relationship" between BPA and diabetes risk, Ning told Reuters Health in an email.
"The evidence was not sufficient enough to declare an association," Ning said.
But an editorial published with the study says the researchers took an "idiosyncratic" approach to their data, essentially dismissing the diabetes risk seen in the highest-BPA group.
"It would be great to see the data released to the wider scientific community," Tamara Galloway, a professor at the University of Exeter in the UK and one of the editorial writers, told Reuters Health in an email.
That's what was done with the US data used in the 2008 study that did find a BPA-diabetes link, noted Galloway - who was one of the researchers on that study.
Another limitation of the study, according to Galloway, is that participants' BPA levels were quite low. The typical level of the entire group was 0.8 nanograms per milliliter, whereas the typical level in the US is closer to 2 nanograms per milliliter.
"The levels in this Chinese population were very low compared with the USA, Europe and Japan," Galloway said, "so the lesson for these more highly exposed populations remain ambiguous."
Ning's team agrees on the need for further studies. Another limit of the current one is that it looked at people at one time point, rather than following them over time to see whether higher BPA levels foretell a higher risk of developing diabetes later on.
The researchers are now following this same study group to answer that question.
Such studies are "crucially needed," Galloway said.
"This is particularly important for associations with cardiovascular disease," she added, "which have been reported now in two separate large-scale population studies."
It's biologically plausible that BPA could affect diabetes risk, according to Galloway.
Studies have suggested that the chemical may act like a hormone in the body, possibly interfering with normal metabolism. Type 2 diabetes arises when the body can no longer properly use insulin, a hormone that regulates blood sugar.
"There are, however, huge gaps in our understanding of the health effects of exposure to this chemical," Galloway said, "We need better data on how people are exposed, and how the compound is metabolised in the body."
For now, she said, people who are concerned about BPA exposure can try cutting down on canned foods and avoiding food containers made of polycarbonate plastics - especially for reheating food, since high heat may transfer small amounts of BPA into food.
Polycarbonate plastics are usually marked with the recycling code "7," Galloway noted.