New data links bisphenol A (BPA) to overall mortality. But the analysis is not clear cut.

Bisphenol A (BPA)

This little fellow is a chemical called bisphenol A, and he’s having a rough time.

From darling of the plastics injury to baby bottle pariah, BPA has generated a lot of controversy over its potentially negative health effects. And the BPA boosters will be brushed back a bit with the publication of this study, appearing in JAMA network open, that links urinary BPA levels to overall mortality in a nationally representative sample.

There is a lot to recommend about this study. Past studies have by and large been cross-sectional — examining BPA levels among people with and without obesity, for instance. But this design makes determining causality really hard. Does BPA make people obese, or do obese people hold on to more BPA?

Researchers, led by Wei Bao at the University of Iowa got around this limitation by using data from the National Health and Nutrition Examination Survey — a federal project that samples people from all over the US collecting data about their… well… health and nutrition — including, in a subset, the concentration of BPA in their urine.

Now, BPA is all over the place.

While it’s hard to find it in baby bottles since 2012 when the FDA banned its use therein, it is still used in canned foods, dental fillings, epoxy resins, and even cash register receipts.

Biologically, we know that BPA has weak affinity for the estrogen receptor — at about 1/1000th of the potency of estradiol. We also know that 90% of Americans have detectable levels of BPA in their urine. But just because you can measure something and it has a chemical name doesn’t mean it’s bad. 100% of Americans have carbonyl diamide in their urine, for example. And if they didn’t this nephrologist would be quite worried.

The researchers divided urinary BPA levels into three groups — low, medium, and high. They found that the rate of death from any cause was about 50% higher in the high group than the low group.

Deaths due to cardiovascular disease were higher, though not statistically significant. There didn’t seem to be an association with cancer deaths.

So… these are confusing results and lead to a number of questions. First of all, out of 3,883 people in the study, 344 died. But only 146 — or 42% — are reported to have died of cardiovascular disease or cancer. That’s weird because, in the US at least, these two causes of death are really the vast majority. The third most common cause of death in the US after cardiovascular disease and cancer is accidents. This gives me some concern about the data quality.

But putting that aside — could this be a causal association? Could BPA be poison? It’s hard to tell.

Perhaps what we’re seeing here is the effect of drinking sugary beverages (from those plastic bottles) or eating a lot of canned food, which may be proxies for lower socioeconomic status or other risk health behaviors.

But there’s another issue here — something as a nephrologist I hold close to my heart.

The problem with measuring things in urine, as opposed to blood, is that urine can be concentrated or dilute. In fact, concentrated human urine is more than 20-fold more concentrated than dilute human urine. (You can feel free to use that fact at your next dinner party).

So you have to account for concentration. If your measurement of urinary BPA was done after the individual drank a bunch of water, you’d get lower BPA concentration. On a hot day when they haven’t been drinking? Higher BPA levels.

The authors realized this and did what a lot of people do in this situation which is to adjust for the concentration of urine creatinine. The idea is that creatinine levels are higher in concentrated urine, lower in dilute urine so by accounting for that, you can take concentration out of the equation. I am not a fan of this approach. Here’s why.

Dilution is only part of what determines creatinine levels in the urine.

No, kidney function has nothing to do with it — that’s blood creatinine. Urine creatinine levels are determined by urine dilution AND muscle mass. If you have higher muscle mass, all else being equal, your urine creatinine concentration is higher.

Now you may see the problem.

People with higher muscle mass will have higher urine creatinines, making their urine (by the logic of this study) appear more concentrated than it is, leading their BPA level to be revised downward. High muscle mass can thus artificially lower their assessment of BPA.

What is high muscle mass associated with? Living longer — for a bunch of reasons including the simple burden of chronic disease.

That is why multiple studies, including this one from yours truly, have shown that lower urine creatinine concentration is a predictor of all-cause mortality.

Cool right?

This is easy to fix though — just adjust for urine osmolarity instead of urine creatinine. I reached out to Dr. Bao to see if we could untangle this but didn’t get a response by deadline.

So, here we are. Is BPA intake going to kill you? I’m not sure. The evidence here is really somewhat… dilute.

A version of this commentary first appeared on

Addendum: Due to concerns raised in this commentary, the study team re-analyzed their results using urine osmolarity, instead of urine creatinine, as a proxy for urine concentration. The adjusted HR was 1.42 (0.95–2.13) for the highest tertile of BPA compared to the lowest.

Writing about medicine, science, statistics, and the abuses thereof. Commentator at Medscape. Associate Professor of Medicine at Yale University.

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store