BPA in canned food, and why two food-safety agencies cannot agree on the dose

1. 01

   Five days of canned soup moved urinary BPA nearly twentyfold

   A randomized single-blind two-by-two crossover trial gave 75 participants one 12-ounce serving of canned or freshly prepared soup daily for 5 days with a 2-day washout. Urinary BPA came out at a geometric mean of 1.1 micrograms per litre, 95 percent CI 0.9 to 1.4, after the fresh arm and 20.8 micrograms per litre, 95 percent CI 17.9 to 24.1, after the canned arm, a 1,221 percent increase and a mean difference of 22.5 micrograms per litre, p below 0.001 (Carwile et al. 2011).

2. 02

   In a normal diet the effect is 24 percent, not 1,221 percent

   Across 7,669 NHANES 2003 to 2008 participants aged 6 and over, eating one canned food in the previous 24 hours was associated with 24 percent higher urinary BPA, 95 percent CI 1.11 to 1.38, and two or more canned foods with 54 percent, 95 percent CI 1.27 to 1.88 (Hartle et al. 2016). The trial figure is what a daily soup habit does; the survey figure is what a normal week does. Both matter, and quoting only the first overstates the ordinary case by a factor of about fifty.

3. 03

   Canned soup is the outlier and canned beverages are not a source

   By category, one or more servings against none was associated with 41 percent higher urinary BPA for canned vegetables and fruit, 95 percent CI 1.23 to 1.63, 70 percent for canned pasta, 95 percent CI 1.18 to 2.44, and 229 percent for canned soup, 95 percent CI 1.22 to 4.30. Canned beverages showed no association (Hartle et al. 2016). Soup is hot-filled, liquid and long-contact; beverage cans use a different coating with cold contents. The ranking is a contact-conditions ranking, not a brand ranking.

4. 04

   EFSA cut its tolerable daily intake 20,000-fold in one revision

   The 2015 temporary tolerable daily intake was 4 micrograms per kg body weight per day. The 2023 re-evaluation identified an effect on Th17 immune cells in mice as the critical effect, set a human equivalent reference point of 8.2 nanograms per kg body weight per day, applied an overall uncertainty factor of 50, and published a TDI of 0.2 nanograms per kg body weight per day (EFSA CEP Panel 2023). Both mean and 95th percentile exposures in all age groups exceeded the new figure by two to three orders of magnitude.

5. 05

   FDA holds the opposite position, and its reasoning is toxicokinetic

   FDA's stated current perspective is that BPA is safe at the current levels occurring in foods. A four-year review completed in 2014 covered more than 300 studies published between November 2009 and July 2013 without prompting a revision, and work at the National Center for Toxicological Research found oral BPA is rapidly metabolised to an inactive form, with internal exposure to the active form predicted below 1 percent of the administered dose and far faster clearance in primates than rodents (FDA).

6. 06

   The 2012 and 2013 FDA amendments were not safety findings

   FDA removed BPA-based materials from baby bottles, sippy cups and infant formula packaging in 2012 and 2013, and states explicitly that those regulatory amendments were made on the ground of abandonment by industry rather than as a safety determination (FDA). This distinction gets lost constantly, because the action reads like a warning and the agency's own description of it does not. Anyone citing the baby bottle rule as proof that FDA considers BPA unsafe is citing the wrong document.

7. 07

   The EU stopped arguing and banned it, coatings included

   Regulation (EU) 2024/3190 prohibits BPA in food contact materials and articles including plastics, coatings, inks and adhesives, and entered into force on 20 January 2025. Transition periods run 18 months for most single-use packaging, to 20 July 2026, and 36 months, to 20 January 2028, for categories including items with BPA-based coatings, which is the can-lining case. Cans on an EU shelf are therefore mid-transition through 2027 rather than already changed.

8. 08

   The EU rule covers other bisphenols, which closes the substitution loophole

   Earlier BPA-free labelling created a well-documented substitution problem: the coating was reformulated with bisphenol S or bisphenol F, the label became accurate, and the structural class did not change. The EU regulation is written against bisphenols as a group rather than BPA alone, so a compliant EU can cannot simply have swapped one letter. A US can labelled BPA-free carries no equivalent guarantee, and the useful question to a manufacturer is which resin replaced the epoxy, not whether BPA is present.

9. 09

   Exposure tracks yesterday, not a lifetime of accumulation

   BPA is glucuronidated rapidly and excreted in urine on a timescale of hours, which is why a 2-day washout was sufficient to reset the crossover trial (Carwile et al. 2011) and why FDA emphasises that internal exposure to the active form is predicted below 1 percent of an oral dose (FDA). The practical consequence is that habitual frequency is the variable worth changing. One can of soup does not enter a long-term body burden; a can a day sustains a twentyfold elevated biomarker for as long as it continues.

10. 10

    Contact conditions decide migration, so format substitution works

    Migration rises with temperature, fat content, acidity and storage duration, which is why hot-filled liquid soup sits at the top of the NHANES category table and cold canned beverages sit at zero (Hartle et al. 2016). It also means a glass jar, a carton or a frozen bag is a genuine substitute rather than a symbolic one. Acidic contents are the highest-value swap, so tomatoes in a jar are worth more than tuna in a pouch if only one change is going to be made.

11. 11

    Rinsing canned food does not remove BPA

    Rinsing removes surface salt and the packing liquid, which is a legitimate sodium intervention and is often confused with a BPA intervention. It is not one. Migration occurs into the food matrix during retort processing and subsequent storage, so the bisphenol is distributed through the contents rather than deposited on them. Draining and rinsing beans is worth doing for the sodium figure on the label, and it should not be counted as reducing bisphenol exposure.

12. 12

    The cheapest fix was already the cheapest option

    The highest-frequency canned item in most kitchens is beans, and the dry version of the same bean costs a fraction of the canned version per cooked cup before any contaminant argument is made. Buying dry, cooking in batch and freezing in portions removes the most frequent exposure and lowers the grocery line at the same time, which is the only intervention here with a positive return. Everything else on the list is a preference, and the rest of the pantry staples argument is unchanged.

13. 13

    Keep canned fish, and be specific about why

    Canned fish is a smaller, less frequent serving than soup and it is doing nutritional work that is genuinely hard to replace at the price, particularly canned wild salmon and sardines for omega-3 and calcium from the softened bones. Nothing in the NHANES category breakdown singles out canned fish, and the format substitutes are poor. Where a canned item is both infrequent and nutritionally load-bearing, the exposure argument does not carry enough weight to displace it.

What people ask us next.

Q01

How much does eating canned food raise BPA exposure?

Two figures apply to two different patterns. A controlled trial feeding one 12-ounce serving of canned soup daily for 5 days raised urinary BPA from a geometric mean of 1.1 to 20.8 micrograms per litre, a 1,221 percent increase (Carwile et al. 2011). Ordinary consumption is far smaller: one canned food in the previous 24 hours was associated with 24 percent higher urinary BPA across 7,669 NHANES participants, and two or more with 54 percent (Hartle et al. 2016).

Q02

Why do EFSA and FDA disagree so badly?

They weighted different evidence. EFSA's 2023 re-evaluation treated an immune effect on Th17 cells in mice as the critical endpoint, set a human equivalent reference point of 8.2 nanograms per kg body weight per day, applied an uncertainty factor of 50 and published a tolerable daily intake of 0.2 nanograms per kg body weight per day (EFSA CEP Panel 2023). FDA weighted toxicokinetics, finding oral BPA rapidly inactivated with under 1 percent reaching the active form internally, and holds that current dietary levels are safe (FDA).

Q03

Is BPA still legal in food cans?

It depends on the market. In the EU, Regulation (EU) 2024/3190 prohibits BPA in food contact materials including coatings and entered into force on 20 January 2025, with transition periods to 20 July 2026 for most single-use packaging and to 20 January 2028 for categories including BPA-coated products. In the United States it remains permitted for can coatings, with FDA maintaining that the current levels in food are safe (FDA).

Q04

Does a BPA-free label mean the can is safe?

It means one specific molecule is absent. The documented substitution pattern replaced BPA with bisphenol S or bisphenol F, structurally similar compounds with far less regulatory scrutiny, which made the label accurate without necessarily changing much. The EU rule is written against bisphenols as a class rather than BPA alone, so EU compliance carries more information than a US marketing claim. The useful question to a manufacturer is which resin system replaced the epoxy.

Q05

Should canned beverages be avoided too?

The data say no. Across 7,669 NHANES participants, canned beverages showed no association with urinary BPA concentrations while canned foods showed clear ones (Hartle et al. 2016). Beverage cans use a different coating system and the contents are cold rather than hot-filled, and migration is strongly temperature-dependent. Extending a canned-food concern to canned drinks costs effort and buys nothing measurable.

Q06

Does rinsing canned beans reduce BPA?

No. Migration happens into the food matrix during retort processing and storage rather than onto its surface, so rinsing cannot remove what has already partitioned into the beans. Rinsing is still worth doing for a different reason, since it removes a meaningful share of the added sodium in the packing liquid. Treat it as a sodium measure and not as a contaminant measure.

Q07

Which canned foods are worth replacing first?

Rank by frequency and by contact conditions. Soup ranks worst on both, at 229 percent higher urinary BPA in the NHANES category breakdown, and it is easy to make in batch. Beans are next by frequency and are the only swap that saves money, since dry beans cost a fraction of canned per cooked cup. Acidic contents such as tomatoes are high-extraction and move well to glass jars. Canned fish is infrequent and nutritionally hard to replace, so it stays.

Q08

Does BPA build up in the body over time?

Not in the way persistent contaminants do. BPA is rapidly glucuronidated and cleared in urine within roughly a day, which is why a 2-day washout was enough to reset the arms of the crossover trial (Carwile et al. 2011) and why FDA emphasises that internal exposure to the active form is predicted to be under 1 percent of an oral dose (FDA). Exposure therefore reflects recent intake, so changing a habit changes the biomarker within days rather than years.