Bisphenol A (BPA) back in the news

We've discussed the component of plastics bisphenol A (BPA) here before (here, here) but yesterday the Journal of the American Medical Association published a significant paper with an accompanying editorial that deserves mention. A panel of the FDA was scheduled to meet the same day to review FDA's draft assessment that BPA was not a safety problem in the US food supply and environment. As a result of the JAMA article, the ranking member of the Committee on Finance, Chuck Grassley (R-Iowa) has written to the Commissioner of the FDA asking for a clarification of the FDA's position on the science underlying a recent NIH review of BPA's safety together with the names, titles, internal communications and communications with the chemical industry trade association and manufacturers of BPA. Clearly Grassley smells a rat. The precincts of the FDA are already so odorous it's a wonder a new stink can even be detected. So this one must really reek. Here's some of the background from the Editorial:

In this issue of JAMA, Lang and colleagues report the results of the first major epidemiologic study to examine the health effects associated with the ubiquitous estrogenic chemical bisphenol A (BPA). This compound is the base chemical (monomer) used to make polycarbonate plastic food and beverage containers, the resin lining of cans, and dental sealants; it also is found in "carbonless" paper used for receipts as well as a wide range of other common household products. Based on their analysis of data from the National Health and Nutrition Examination Survey 2003-2004, Lang et al report a significant relationship between urine concentrations of BPA and cardiovascular disease, type 2 diabetes, and liver-enzyme abnormalities in a representative sample of the adult US population. (vom Saal, F, Myers P in JAMA)

BPA is one of the highest volume chemicals in production worldwide. It has become ubiquitous in the environment, including food and drink, and BPA levels can now be detected in 90% of Americans. BPA also acts like a mimic of the major female sex hormone family, the estrogens. A very large body of animal literature has showed two things: that BPA has biological effects at levels relevant to environmental exposure and similar to natural estrogens (nanomole range); and that in animals it acts through response mechanisms also present in humans. The Lang et al. paper therefore is not surprising. It is pretty much what we would expect to see. On the other hand, the paper is surprising, because looking a representative sample of the US population for common diseases usually involves so much noise you can only see the strongest signals through the static.

The study involved 1455 participants in one of the two year random samples of the US population (NHANES, the National Health and Nutrition Examination Survey) done by the National Center for Health Statistics. Each was asked if a doctor had ever told them whether they had one of a series of conditions, including various kinds of heart conditions, asthma, diabetes, etc. A spot urine sample was taken and a blood draw and further questions about age, etc. This is the dataset used by the researchers based at the University of Exeter (UK) and published yesterday.

This was a very carefully done study but the results are best characterized as preliminary. The adult onset diabetes (Type II) and the liver function results are consistent with existing animal data.

The cardiovascular results (angina, heart attack, coronary artery disease) were new. They have not been looked for in animal models. Considerable pains were taken to evaluate and if possible eliminate alternative explanations but no single study can cover all possibilities. The cross-sectional nature of the design (meaning that the putative cause, BPA exposure, and the disease outcomes, were measured simultaneously, so the temporal sequence that is one hallmark of causal associations could not be evaluated. This will require some well designed prospective studies which should include pregnant women, infants, children and adults. The NHANES data in this study only included adults.

One of the important features of the existing science as underlined in this paper is that the FDA and EPA still base their safety assessments on old, high dose laboratory studies. One thing abundantly clear about hormones is that their effects are not necessarily increasing with dose. Higher doses often have lower effects than lower doses, that is the dose - response curve is U-shaped. Unfortunately the Bush FDA and EPA approved methods only look at the right hand branch of the U and take the bottom as the acceptable exposure. This isn't 21st century biology, as the accompanying Editorial notes. Canada has already called BPA a toxic chemical, not a safe one. It isn't that the Canadians are ahead of the curve. It is that the Bush FDA is behind the curve.

Way behind.

More like this

In your post, the fourth paragraph from the bottom seems to have accidentally been cut off in the middle.

BPA has been associated with ...

I have a question that is rather off topic but

"it also is found in "carbonless" paper used for receipts"

What type of receipts, why would they need to be paperless & why BPA - normally used for hard clear plastics - suitable?

"What type of receipts, why would they need to be paperless & why BPA - normally used for hard clear plastics - suitable?"

You know those triplicate forms: top sheet white, middle is yellow, bottom is pink. Just like the older carbon-copy system where you had a black carbon sheet inbetween the useful layers. Apparently BPA is part of the chemical treatment for the yellow and pink sheets, turning it color when pressure is applied.

Thank you Kevin

I read it, it seemed strange and it was going to bug me if I didn't ask.

Stuart: Thanks. It was actually a fragment of a deleted sentence. I've corrected it.

how can I test my bisphenol-A level ?

anon; It requires a specialized lab. The paper used analyses performed with solid phase extraction coupled with HPLC (high performance liquid chromatography). - isotope dilution tandem mass-spectrometry with peak focusing (according to the paper). So this is a research technique at the moment.

If you want to identify the questionable bottles folks then you look on the bottom. If there is a recycle symbol and a 7 on the inside of it then its BPA positive. The bottles of Listerine, drinking water in some cases, lots of drinks in the larger variety have it. Another test is the rigidity of the bottle. If its hard walled then its likely got the BPA in it.

Revere-The EU hasnt bailed on this yet, even though the Canucks have. There are questions of course but its not just the Bush FDA. I would though like you err on the side of caution. The plastics people of course resisted but I bet they are scrambling else they'll get sued for everything from birth defects to girls breast reductions.

I use glass because I dont want all that crap sitting out on the side of the road. Little more dangerous but I can wash that stuff out.

By M. Randolph Kruger (not verified) on 17 Sep 2008 #permalink

This study is interesting. However I suspect that the observed association may be more an effect of metabolic disturbances than caused by BPA. In all cases the standard deviation of those with a named disease was higher than those without (in the figure). In all cases (except for liver disease) the BPA levels were higher, in the case of liver disease it was lower. BPA is metabolized by the liver, which attaches a glucose molecule to it to make is water soluble so it is then excreted in the urine. Liver disease might reduce that and increase excretion in the bile. Are we seeing differences in partitioning of BPA due to the underlying diseases independent of any cause by BPA?

The absolute differences in BPA levels were small. They stratified on the basis of percentiles. There is a pretty clear effect of BMI on BPA levels with the highest levels at the highest BMI. Since most of these diseases are also associated with BMI, that could be a confounding factor. Similarly there was an increase in non-Hispanic black.

The main exposure route is thought to be consumption of individually packaged beverages. Could the urinary BPA levels they measured simply be a surrogate for consumption of canned soda? That might explain the connection to BMI.

When molecules are used as signaling compounds in a feedback metabolic pathway, there is no threshold for effects. The reason is because the feedback system is already operating in the active range that is acting where an increase or decrease in the signaling molecule transmits a signal and that signal causes physiology to respond. That is how all feedback systems work. If you affect the components of the feedback system, there is no threshold for changing the outcome of the system.

The no threshold effect is true for all feedback systems. It is true for hormone mediated feedback systems, it is also true for nitric oxide mediated feedback systems. Because steroids are produced and processed by the cytochrome P450 system, and that system is regulated by NO, a change in the basal NO level also has an effect on the outcome of that steroid mediated feedback system with no threshold. Estrogen activates the estrogen receptor which then activates nitric oxide synthase and makes NO. That feedback in an important part of estrogen mediated physiological effects. That is part of how estrogen improves bone density, the NO from estrogen stimulates osteoblasts to deposit more bone mineral and inhibits osteoclasts preventing resorption.

Changing the NO level can have the same effects as changing the estrogen level. Organic nitrates have equivalent effect in preventing osteoporosis in post-menopausal women as does hormone replacement therapy. I think that it is low NO (from the bacteria I am studying) that causes some of the hormone disruption observed in things such as polycystic ovarian syndrome. Low NO is a stress response, make NO low and physiology responds as if it is under more stress. Chronic stress is a bad time to get pregnant, so physiology causes infertility, higher testosterone to make body size bigger, and more hair to provide a bigger niche for these bacteria.

Another take you might be interested in.

One point that derives from Steve Novella's take is that that Europe must be even further "behind the curve,"at least by your definition. Another is that the decision of EU's regulatory apparatus not to take action on bisphenol A can hardly be tarred as being manipulated by the Bush Administration.

One thing frustrating thing about BPA is that, should you choose, based on evidence such as this, not to use it... it's really hard to eliminate. I thought I had done just that by ditching my Nalgenes (from which I'd drunk rivers of water in college) and switching to a stainless steel water bottle. And then I found out it's in the lining of cans, and is mostly likely to leach with acidic foods such as tomatoes. Problem is, alternatives such as home canned tomatoes are five times as expensive. (And canning your own a huge commitment of time, energy, and money.)

They talk about BPA in commercial products, but I haven't heard much about them in medical devices. Is it used in, say, feeding tubes?

"On the other hand, the paper is surprising, because looking a representative sample of the US population for common diseases usually involves so much noise you can only see the strongest signals through the static."

Can you explain this statement further? What "noise/static" are you referring to?

daedalus2u- or others who know their BPA facts-
There are a lot of businesses that involve casting or laminating epoxy resins. That often means a guy in jeans and a t-shirt (but wearing a week old $15 organic respirator mask and gloves) mixing up a tub of ~50:50 bisphenol A and epichlorohydrin and then pouring, brushing, or squeegeeing the goop on a boat or bartop or whatnot. After that the hardened resin is often sanded and polished by another guy wearing a particulate mask ...

1. Has anyone measured bisphenol A levels in boat construction/repair workers, surfboard glassers, or other people who do hands-on work with epoxy?

2. How quickly is bisphenol cleared from the system of a (non-obese, healthy-livered) person? Days, weeks, months?

Kathryn -- I think the idea is that a representative sample of a country's adult population encompasses such a wide variety of people (by age, gender, etc.), that in order to be able to pull out a correlation like this, the correlation has to be very strong.

Because very few of the youngest adults would have been diagnosed with coronary artery disease, their presence in the study depresses the overall proportion of those reporting coronary artery disease... while they might well have just as much BPA in their urine older people in the study.

The fact that they can still find a correlation between BPA and coronary artery disease when they include 18-year-olds in the study would lead a reasonable person to hypothesize that if they narrowed the study to those who are more likely to have coronary artery disease (say, men over fifty and post-menopausal women), they might find an even stronger correlation. Same goes for adult onset diabetes; plenty of twenty-year-olds have it, but not as many as forty-year-olds or sixty-year-olds.

caia, you may be right, but if BPA correlates with obesity (which it appears to from the data in the paper), is that cause, or effect? The major dietary source is said to be canned beverages. If obese people, or people with the metabolic syndrome drink more canned soda (regular or diet), that could result in the BPA association simply from dietary choice.

According to one of the papers it said that BPA is metabolized in the liver by adding a sugar molecule to it (so it is more easily excreted) and once that is done, BPA+sugar doesn't activate the estrogen receptor. But some BPA+sugar gets excreted in the bile, where the sugar gets cleaved off and the BPA can be reabsorbed (and go through the liver again). BPA+sugar is excreted in the urine in humans (not so much in rodents). These were spot urine samples. If the time constant for conversion to BPA+sugar is very fast (which is seems to be considering they are talking about "first pass" conversion (that is conversion the first time it goes through the liver), then the excretion might be quite rapid and so a spot urine test would mostly reflect very recent BPA consumption (since the last urination). There are also other metabolites with different activity profiles.

These animal studies with low level steroid mimetics are difficult to do. The gigantic problem is that the background of steroid mimetics is not zero. The background in food may greatly exceed the level in the agent you are trying to test. There are many hormonally active agents in virtually every type of food. The diet of the experimental animals can affect things, the composition of their cages, even how the cages are cleaned. One of the early studies was on how using polycarbonate cages made a big difference that was eventually traced to BPA from the polycarbonate. That was found because an earlier experiment was screwed up by not using identical cages. Even the age of the cages makes a difference, what other plastics were processed on the plastic processing equipment the manufacturer used to make them. The containers used to hold the food, and water given to the animals matters. The levels of phytoestrogens in food are variable depending on how the food is grown, what specific strains, the level of stress the plants were under.

An experimenter may try to change "one thing", but changing "one thing" may affect dozens or hundreds of things in the test organism. There are many different steroid receptors, with different sensitivity to different compounds, in different tissue compartments doing different things at different times. This paper shows some of the backgrounds of "natural" endocrine disruptors are not small, and probably exceeds endocrine disruption effects from synthetic pesticides by a couple of orders of magnitude.…

The big problem in studying compounds that mimic signaling compounds in the body is that the body uses them at different concentrations for all different kinds of things. High levels for very local signals inside a cell, somewhat lower levels to signal inside an organ, still lower levels to signal from one organ to another. The brain makes a lot of steroids and uses them to regulate things everywhere else in the body. Lots of different organs make the same signals that get added together and read out by still other organs depending on things that we don't know a lot about. The dynamic range of steroid levels is quite a few orders of magnitude (I am talking dynamic range between different tissue compartments, not dynamic range in plasma (which isn't the most important compartment for steroids anyway, but is the only one that can be measured)).

The local levels in the organs that make these steroids can be very high. With steroids signaling different things at different concentrations, different concentrations of BPA will screw up different pathways by different amounts in different tissue compartments, and it can screw it up in different directions. BPA does activate some receptors, it also inhibits some enzymes that metabolize steroids, so BPA could act like the steroid, activate the receptor and cause feedback inhibition of the steroid so levels go down. If it inhibited the rate limiting enzyme that removes it, the level could go up. It could do either in different tissue compartments. Even if you are only looking at BPA, you are looking at it in competition with all of the other compounds that the receptors, enzymes, transporters, etc. have any affinity to.

The most sensitive time by far is the first trimester when the major differentiation is occurring in the fetus. It is during differentiation that most of the epigenetic programming of physiology is done. It is the epigenetic programming during differentiation that sets the differences between the different cells. The "only" difference between a liver cell and a brain cell is their different epigenetic programming. Steroids are important in regulating those differences. For example, it is mostly (only?) steroid signaling that determine gender. A person with androgen insensitivity is (mostly) anatomically female even if they are XY in their chromosomes. Reptiles have no sex chromosomes. Gender is solely determined during development, usually by temperature. Adding the right steroid at the right time will override that temperature signal in reptiles.

Exposure to any endocrine disrupting agent is likely bad during the first trimester. That may be why many women become nauseous in early pregnancy to avoid eating, and so avoid exposure to potential endocrine disrupting agents in food. Metabolic needs are not increased during the first trimester. What is most important is proper differentiation of the fetus, expansion of the placenta to support greater metabolic needs later, expansion of the maternal blood and mitochondria supply to do the same thing. Hematopoiesis is triggered by NO, as is mitochondria biogenesis, as is angiogenesis. First trimester is a high NO state, and I think that high NO also induces nausea. There is a positive correlation with nausea and pregnancy outcome, that is women who were nauseous had better pregnancy outcomes than women who were not. Nausea may be an unpleasant side effect of a desirable high level of NO which has the desirable effect of limiting consumption of endocrine disrupting agents during the most vulnerable time of the life cycle.

daedalus: The paper is a little unclear as to how the reverse causation issue was handled. The authors recognize this as a possibility but the data on logged BPA adjusted for age, sex and creatinine across categories isn't shown but it is stated to be not significant. I'm not sure exactly what they are saying there. Still, you suggest one of the many alternative possibilities. The rather extraordinary explosion in obesity requires some explanation besides inactivity and fast food IMO, so these data are certainly of interest. It's like a sign in the ground that says, "Dig here." Dig we must.

Revere, I completely agree that obesity requires explanations other than inactivity and diet. It seems pretty clear it has to be something that perturbs the feedback loops that physiology uses to regulate those things. That perturbation has to be from something that is working inside those feedback loops.

I dont think that those nuances of control theory are appreciated by people working on obesity causes. For weight to be stable long term, physiology has to be working to the right setpoint. If the setpoint is biased, physiology will call for more or less calories until the now biased setpoint is reached. If there is enough bias that may never happen. I think that is the only mechanism by which people can become hundreds of kg over weight. I think it is a similar bias (but in the opposite direction) that causes anorexia.

The more I get into NO physiology, the more NO physiology just screams at me that it is involved. Of course you have to get pretty far down in the nitty gritty details to have that perspective. Down where you get into where there is so much detail it is hard to see the forest for the trees.

I think you should look at this study from the opposite point of view. The researchers took publicly available data and ran a few correlations or regressions on it. Since something significant turned up, they were able to publish it in JAMA. If nothing had been significant, they probably wouldn't have published it at all. This could easily just be a chance finding. And the only health outcome seems to be reported past history of heart disease and other conditions, as compared with their current levels of bisphenol A. So we don't know if their current BPA levels have anything to do with their past history and the temporal sequence is reversed from what the implications suggest. It is implied that BPA leads to heart disease, but we have no idea what their BPA levels were when they got heart disease. In fact we really dont even know if these medical histories are particularly accurate. This is really more of a hypothesis-generating study than anything else.

Carolyn: You raise some well worn issues regarding multiple comparisons which are too complicated to hash out here. Except for the cardiovascular issues, however, these are effects consistent with a rather extensive body of laboratory data on BPA. Moreover there were a number of subsidiary analyses done to tease out spurious associations. The public availability of the data is irrelevant. Every mortality study is done with publicly available data, too. And if the number of comparisons is few, as you allege, then there is no multiple comparisons problem.

This is clearly billed as preliminary data but quite striking preliminary data in light of the large amount of supporting animal data. So it is just as plausible to call this a hypothesis confirming study. But this is, as I say, a pretty tangled area and not settleable within the format of a comment.

The Lack of Efficacy of the FDA

The Food and Drug Administration (FDA) originated in its original form several decades ago, with thier objective being to ensure the health and safety of the citizens of the United States regarding products that they consume, and the catalyst for its development was due to the concepts originated by a man named Upton Sinclair. However, the objective of the FDA seems to have changed, as they appear to have formed a pathological alliance with the pharmaceutical industry possibly through the money that this industry gives the FDA for various reasons, so it seems. This is such a large amount of funds issued to the FDA by the industry that it has resulted in possibly half of the FDAs annual income for various reasons. Results of this relationship, one could posit, have been the approval of unsafe drugs on occasion and lack of sufficient regulation and monitoring of the pharmaceutical industry that the FDA is obligated to perform. In addition, the FDA lacks sufficient resources to complete thier duties completely, it has been said. In fact, in the FDA's own mission statement, they claim that they will always protect, advance, and improve public health.

An example of this support from the pharmaceutical industry is the Prescription Drug User Fee Act, which requires drug companies to pay around a million dollars for a priority review of a potentially approved drug of theirs. Implemented by the FDA in 1992, it has required increases in the amounts paid by drug companies since then. In fact, of the FDA's estimated 2 billion dollar budget, about 20 percent of this money is from this act. Since 1992 and the implementation of this act, there have been no less than 12 drug recalls, which statistically is significant. To improve public health, it may be best that the FDA receive all funding from the U.S. Treasury.

So this intimacy between the FDA and pharmaceutical industry seems to continue to progress, as illustrated with the new proposal by the FDA to allow the pharmaceutical industry's reps to discuss their products with prescribers off-label, which means that the FDA somehow is accepting and allowing the sales reps of the drug industry to possibly create harm to patients with this proposal due to uncertainty associated with unapproved uses of a drug promoted in this way, many believe.

A prescriber, upon their own discretion, can in fact prescribe a drug off-label. Historically, however, ad for good reasons, representatives from the pharmaceutical industry have been prohibited from suggesting or discussing off-label possibilities with thier promoted products with prescribers, due to the speculative nature of the concept. In fact, it is a federal offense for such reps to speak off-label about the drugs they promote, and some have been penalized for this behavior in the past in the form of monetary settlements by the Department of Justice, yet appears such penalties are not much of a deterrent for this behavior that may now be authorized and has continued to occur regardless of possible future permission by the FDA.

This FDA protocol that is being considered, called, Good Reprint Practices, would require reps to use what in fact may not truly exist, which is truthful and authentic clinical trials related to off-label possibilities of thier promoted products during any dialogues they may have with prescribers. This in itself may lack a necessary degree of validity for such discussions by reps, as it has been suspected that some clinical trials are flawed due to the trials being possibly manipulated by the pharmaceutical companies of the meds involved in such trials through possibly third parties, such as trial deception involving ghostwriting and invalid authors of such trials, which disintegrates the quality of such a study. These facts can be validated and have been discovered by others, so caution instead of autonomy should be utilized regarding clinical trial discussions with health care providers, some may say.

Furthermore, this proposal is additionally flawed due to the fact that most pharmaceutical reps have little of any clinical training. So the ability for pharma reps to analyze data from trials justifying an off-label concept is unlikely. This complicates the idea of this off-label concept- this ignorance of most drug reps in regards to the complexities of these once reliable and dependable methods of proof, and that is the clinical trials. In addition, the proposed relaxation of previous restrictions regarding off-label promotion could prove to be a catalyst for reps to manipulate statements to prescribers for their own benefit in regards to their promoted meds while disregarding the health of the consumers of these meds. So, our previous safety administration, the FDA, appears to be evolving into taking somewhat of an apathetic stance regarding the safety of public health by suggesting such practices with what appears to be deliberate intent and reckless disregard for public health, so it seems. This may be due to the FDA now viewing the pharmaceutical industry as thier sponsor or client. Its unbelievable this proposal ever came into existence, with the delusional fallacy that it would be an actual benefit to public health. Furthermore, this FDA proposal may complicate existing patient medication errors, such as in the elderly or dosing for children, complicated by the fact that many are unable to understand label instructions on their med. So there are enough problems with prescribing medications correctly, and adding this FDA proposal would just make the situation worse, it seems.

However, there is freedom of speech. Perhaps another alternative would be to have clinically trained people discuss such issues with prescribers, instead of the pharmaceutical sales reps, who, unlike those who may be more academically enriched, have the sole objective of increasing the market share of their promoted meds. Such people, ideally, would not have any association with the makers of such drugs spoken about to prescribers, yet this may be unlikely to occur. Regardless, awareness needs to be achieved by the public about the possible dangers of this FDA proposal, even though it's possible that the public has heard about it through media sources. As citizens, we have the right to insist that the FDA demand and direct the pharmaceutical company to prevent potential harm that may come to patients in need instead of thier apparent desire to please thier bread and butter. More congressional oversight is needed of this Association.

"It is difficult to get a man to understand something when his salary depends upon his not understanding it." --- Upton Sinclair

Dan Abshear

Author's note: what has been written is based upon information and belief