
[Adobe Stock]
How a recycling study spawned spatula hysteria
The study causing spatulageddon is “From e-waste to living space: Flame retardants contaminating household items add to concern about plastic recycling” published in the journal Chemosphere. The corresponding author is affiliated with Toxic-Free Future. The study found flame retardants where they shouldn’t be, in objects used to touch food. The graphical abstract stated 17 in 20 items contained flame retardants. It also stated flame retardant chemicals were found at levels up to 22,790 mg/kg in food serviceware, hair accessories, kitchen utensils and toys. A now banned compound, BDE-209, was deemed particularly worrisome. “Estimation of exposure to BDE-209 from contaminated kitchen utensils indicated users would have a median intake of 34,700 ng/day, exceeding estimates for intake from dust and diet” according to the authors. They went on to conclude “Products found in this study to contain hazardous flame retardants included items with high exposure potential, including food-contact items as well as toys.” Pretty alarming stuff. A PR push by Toxic-Free Future resulted in headlines like “Black Plastic Kitchen Tools Might Expose You to Toxic Chemicals. Here’s What to Use Instead” from The New York Times and “Throw Out Your Black Plastic Spatula” from The Atlantic. Spatulageddon was on.

Chart based on data from Liu et al. (Chemosphere 2024) and analysis in McGill University’s Office for Science and Society article
[Also see: Microplastics are bad, but ignoring science is worse]
The flame retardant boogeyman
Finding flame retardants in recycled plastics, especially ABS, is not new. It is understandable. ABS is a common polymer for electronics like TVs, monitors, computers, keyboards, mice, phones and such. It is in things we have in our homes and touch every day. We demand recycling. Some contamination is inevitable. It also is inevitable that some replaced materials will show up since the ABS is going into durable things getting recycled years, maybe decades, after production. Papers going back to at least 2013 have measured brominated flame retardants in recycled ABS. That 2013 work also found elevated levels in spatulas, but received very little publicity. Studies in the intervening years reinforced that hazardous additives lurked in recycled polymers, again receiving little notice. Toxic-Free Future got more publicity this time around. It was impossible to miss the reporting but it seemed incomplete. I have toured MRF facilities and heard how hard it is to recycle black plastic. I found myself wondering about black food contact items, but wasn’t focused on spatulas. Many of the take-out containers are black with clear lids. Most I’ve encountered are polypropylene, some are PET according to the recycling symbols. Some indicate they are microwavable. I don’t like heating in plastic and like it less if I know the plastic is recycled. Bad actors that stay put normally are far more likely to migrate at higher temperatures. Wanting to know about the black containers I encounter, I got the article and studied it and the supplemental data. It didn’t directly address my concerns, but did indicate only very low levels of flame retardants in all non-styrenic polymers. ABS and high-impact polystyrene (HIPS) were high, polypropylene, PET and nylon were low. It isn’t damning of the takeout containers I was questioning, but still left lots unanswered. Some inconsistencies got me making spreadsheets, checking calculations.
There are a couple of points worth making. The first is reiterating flame retardants are in objects we already have and touch on a daily basis. I have no doubt that the keys I’m touching as I type this have flame retardants in them. Testing indicates that normal touching does not create an unreasonable risk. It wouldn’t be true if I started chewing on my keyboard and mouse. It wouldn’t be true if I started cooking on them, unlikely as that is. The second point is that dilution occurs in recycling. The levels aren’t amplified above what was present in the materials being recycled. Recycling will always be an activity that plays the odds. It is impossible to test for all materials that could potentially cause harm. The risk is managed through dilution keeping the probability of exposures at concerning levels low. Lastly, we have to rely on limits set based on testing and either through voluntary or mandated safety limits. These limits typically are set erring on the side of safety. That said, people can disagree on the magnitude of the hazard and also on what constitutes reasonable risk. Over time, views can change, as they did for BDE-209. It was replaced but still shows up in recycling just by the nature of recycling.
How a decimal point derailed dinner prep
Early in 2024, the Freakonomics podcast introduced me to the concept of convenient errors. Leif Nelson of Data Colada explained that a one decimal point error, potentially a typo, may well escape notice if it is the direction supporting the hypothesis. That is exactly the type of error that damned the spatulas. Authors missed an easily detectable factor of 10 error. Reviewers missed it. Editors missed it. Once stories started to appear, sharp-eyed readers found it. A correction is now affixed to the electronic journal version of the paper.
Digging into the paper left me disheartened. To determine whether BDE-209 presented an acceptable or unacceptable risk, the EPA reference dose of 7,000 ng/kg of body was used. A reference dose is the maximum amount that will not cause harm. 7,000 ng/kg body weight per day for a typical 60-kg adult is 420,000 ng/day. A convenient error was made. 42,000 ng/day was mistakenly used as the reference dose. Estimated exposure was still below the reference dose even with the error. Instead of concluding levels below the reference dose were OK, 80% was reported as too close for comfort and a reason for concern. Once corrected, the exposure estimate drops to a more comfortable 8% of the reference dose. Spatulas might have been saved had it not been for the convenient error.
When half-baked data cooks the books
The estimated exposures are well below the corrected reference dose. The correction now attached to the article details the error but does not back away from the conclusion that recycled plastic in food contact applications is an unreasonable risk. Others have questioned how the conclusions remain valid.
I am doubtful the paper would have been noticed with the correct reference dose. A headline claiming levels claiming spatulas are OK wouldn’t garner much attention. The convenient error is likely the only reason the work got noticed.
The convenient error is only the tip of the iceberg. There is even more reason to doubt the conclusions. Data were cherry-picked, ignoring most of the 203 samples collected and screened. Only 20 objects testing above 50 ppm bromine were analyzed for flame retardants. 183 items didn’t have concerning levels. The ambiguously worded abstract states flame retardants were found in 85% of the samples when, in reality, they were detected in only 17 of 203 samples, only about 8%. The ratio is similar for kitchen items. Only 9 of 109 measured high enough for additional testing. Rather than 85% being suspect, over 90% were OK. The graphical abstract stating that 17 out of 20 contained flame retardants was certainly misleading.
The tall corgi fallacy (or, how to spin a statistic until it barks louder than it bites)

[Adobe Stock]
Levels present in the items does not guarantee an exposure. Correlations between concentrations in the plastic, taking into account use, are needed. Luckily others had already done that work. Previous work by other researchers formed the basis for exposure estimates based on exposure to hot oil. That work concluded handling of objects was unlikely to cause exposure, yet that correlation was used irrespective of how the items would be used. In particular, a relationship connecting BDE-209 concentration to exposure upon heating in oil was used in the recent work. It was used whether the use pattern would ever expose the item to hot oil during cooking. Spatulas, called slotted turners in the paper, are the only items where hot oil exposure and subsequent consumption of food cooked in the oil was likely. Other kitchen accessories, like peelers, never experience hot oil. Food serviceware, hair accessories and toys never see hot oil. Yet, all of the 20 samples where measurements were made of flame retardant levels, including food serviceware, toys, hair accessories and kitchen items like peelers were treated with the same correlation, one developed for the specific case of hot oil emersion and subsequent consumption of the oil. 34,700 ng/day is calculated using the average of all 20 of the BDE-209 measurements. The highest measured spatula gives an exposure estimate less than 0.4% of the EPA reference dose. Spatulageddon should never have happened.
Why spatulageddon was a recipe for overreaction
A couple of things are clear. The spatula killing study is flawed. It is a peer reviewed study where an obvious error still got published. It illuminates failures of peer-review. Convenient errors slip through even when obvious. More subtle errors are surely being missed, as they were here. Once ominous sounding science hits the press, getting it corrected in public opinion is unlikely. Recycling does create risks that aren’t present with virgin material. We have to trust actors in the recycle chain are keeping those risks tolerable. This study reinforces we are recycling with tolerable risks. Lastly, if you didn’t throw out your black spatula, don’t. Spatulageddon was an overreaction.
Mark E. Jones, Ph.D., retired in 2021 as Dow Chemical’s Executive External Strategy and Communications Fellow, following a 30+ year career spanning catalysis, renewable chemistry, and sustainability innovation. Now Creative Director of MJPhD, LLC, he consults, writes, and speaks on science communication, serves on national committees, and judges prestigious industry awards while remaining active in organizations like the American Chemical Society (ACS Fellow) and National Academies.