In this article
- What sodium benzoate is
- Why manufacturers use it
- Where you will find it
- The benzene question: when a preservative meets vitamin C
- Hyperactivity: the Southampton trial, and what it did not test
- Allergy and intolerance: real, uncommon, and hard to prove
- What regulators say
- The debate
- Who should avoid it
- How to spot it on a label
- Alternatives: what replaces it
- Frequently asked questions
Sodium benzoate (E211) is the most widely used food preservative in the world, and one of the most studied: it is generally recognised as safe in the US, permitted in the EU, Australia and New Zealand, and Canada, with an EFSA intake limit of 5 mg per kilo of body weight a day — and a WHO committee limit raised to 20 mg/kg in 2021, four times higher. Its two real questions are specific, not general. It can form trace amounts of benzene when a drink also contains vitamin C and is stored warm or in light — the 2005–2007 surveys found a handful of drinks over the drinking-water limit, and those products were reformulated — and it was one ingredient in the colour-plus-benzoate mixtures that slightly raised hyperactivity scores in the Southampton trial, though no study has ever tested benzoate alone. Allergy is real but uncommon, and usually a pseudo-allergy concentrated in aspirin-sensitive people. For most people at normal intakes it sits well inside every regulator's limit; the exposure flag is for toddlers and children who drink a lot of flavoured drinks.
What sodium benzoate is
Sodium benzoate is the sodium salt of benzoic acid — a white, odourless powder that food and cosmetic makers use to stop yeasts, moulds and some bacteria from growing. The US Food and Drug Administration's definition (21 CFR 184.1733) calls it "the chemical benzoate of soda (C7H5NaO2), produced by the neutralization of benzoic acid with sodium bicarbonate, sodium carbonate, or sodium hydroxide." Its CAS number is 532-32-1, its E number is E211 (INS 211), and its molecular weight is 144.11.
It works only in acidic foods — roughly pH below 4.5 — because in acid conditions it converts to undissociated benzoic acid, the form that actually suppresses microbes. That is why it turns up in soft drinks, pickles and dressings rather than in milk or bread: the acidity is part of how it works, not an accident of the recipe.
Benzoic acid itself is not purely synthetic. It occurs naturally in cranberries, plums, apples and cinnamon, and a Taiwan health-authority monograph puts natural benzoic-acid levels in foods at 10–1,000 mg/kg, including blueberries, coffee beans, green peas, tea and milk. The preservative on a label and the compound in a cranberry are the same molecule; the difference is the dose and the company it keeps.
One conversion matters for everything that follows. Sodium benzoate and benzoic acid are not the same weight: 100 mg of sodium benzoate is equivalent to 84.7 mg of benzoic acid. Regulators express their intake limits as benzoic acid, so a limit of 5 mg per kilo of body weight a day is about 5.9 mg per kilo as sodium benzoate. Keep that 84.7% in mind whenever a number in this article is quoted "as benzoic acid".
Tip: E211 on a European label and "211" on an Australian or New Zealand label are the same ingredient as "sodium benzoate" on an American one — three numbering systems, one preservative.
Why manufacturers use it
Sodium benzoate is popular for the same reason it is hard to notice: it does an unglamorous job cheaply, at low doses, without adding flavour, colour or sweetness of its own.
- It stops spoilage. As an antimicrobial agent it suppresses yeasts, moulds and some bacteria — the organisms that make an opened bottle of juice go fizzy, sour or mouldy. In acidic foods it is one of the most effective preservatives available at the dose used.
- It is cheap and stable. A small amount protects a whole batch, and the ingredient itself keeps well in storage — which is why it is the default preservative in soft drinks, dressings and sauces worldwide.
- It has a second, minor job. The FDA's GRAS entry also lists it as a flavouring agent and adjuvant, and the FDA's Substances Added to Food inventory lists further technical effects including antioxidant, colour/colouring adjunct, flavour enhancer and pH control agent.
- It works beyond food. The same preservative action is used in cosmetics and personal care — hair products, baby wipes, toothpaste and mouthwash — and in pharmaceuticals.
None of these uses is secret or deceptive in itself. The reason sodium benzoate lands on a flagged-ingredients list is not that manufacturers hide it — it is printed on the label in plain sight — but two specific questions about what it does once it is in the bottle: benzene, and behaviour. Those are the next two sections.

Where you will find it
Sodium benzoate is one of the most widely used food ingredients in the world, and it turns up in places its name would never suggest:
- Carbonated drinks and fruit juices — the classic home of benzoate, and the products at the centre of both the benzene surveys and the hyperactivity trial
- Salad dressings, sauces and condiments, including pickles
- Jams, jellies and marmalades, and frozen yogurt toppings
- Hair products, baby wipes, toothpaste and mouthwash
- Some medicines, where it acts as a preservative
It is declared under its own name. In the US it must be declared as "sodium benzoate" in the ingredients list, in descending order of weight. In the EU it appears as "sodium benzoate" or "E211"; in Australia and New Zealand it appears as "preservative (211)" or "sodium benzoate (211)". If it is near the top of the list, there is a lot of it; in a soft drink it is usually present at a fraction of a percent.
The products that matter most for this article are the drinks. A flavoured drink can contain both of the ingredients in the benzene story — benzoate and vitamin C — and it is also the product category where children's exposure is highest. The sweetener in many of the same drinks is high fructose corn syrup, which carries its own dose-and-comparator debate — but the two ingredients are separate questions, and this article deals only with the preservative.
The benzene question: when a preservative meets vitamin C
This is the central concern about sodium benzoate, and it is real chemistry, not a rumour. In 1993, Gardner and Lawrence showed that benzoic acid can be decarboxylated to benzene when ascorbic acid (vitamin C) and a transition-metal catalyst — copper or iron, present in trace amounts in water — are also present. Ascorbic acid reduces the metal ions; the metal-catalysed reduction of oxygen and hydrogen peroxide generates hydroxyl radicals; the hydroxyl radical attacks benzoic acid and knocks off its carboxyl group, leaving benzene. The authors concluded that benzene can be produced "under conditions prevalent in many foods and beverages" and that, since benzene is carcinogenic, "its potential formation in foods during processing and storage should be of some concern."
The details matter. In the 1993 model solutions, benzene formation was maximal at pH 2 — about 37 nanomoles per litre, roughly 2.9 micrograms per litre — and significantly lower at higher pH. Heat, light (UV) and storage time all increase formation, and metal ions catalyse it. Germany's Federal Institute for Risk Assessment (BfR), reviewing the work in 2005, added a caution that belongs in any honest account: "Whether and to what extent benzene is actually formed in the corresponding foods cannot, however, be reliably assessed on the basis of the data available." Laboratory tests prove formation under certain conditions; how much forms in a real bottle was not reliably assessable then.
The history runs in three acts. The FDA first became aware of benzene in soft drinks in 1990, alerted by Cadbury Schweppes and Koala Springs; it tested more in 1993 and found no problem at the time. The issue resurfaced in November 2005, when private citizens' test results reached the FDA — and this time the surveys that followed found real, if small, numbers.
The 2005–2007 surveys — what was actually found:
- United States (FDA, 2005–May 2007, almost 200 samples). "A small number of products sampled contained more than 5 ppb of benzene" — ten products in total above 5 ppb. Manufacturers reformulated those products, and the re-tested products came in under 1.5 ppb. "Almost all samples analysed contained either no benzene or levels below 5 ppb." The FDA's conclusion: "the levels of benzene found in beverages to date do not pose a safety concern for consumers."
- United Kingdom (Food Standards Agency, March 2006, 150 beverages). 43 contained benzene; four contained levels above the WHO drinking-water standard of 10 ppb and were withdrawn from sale. The other 146 were not detectable or between 1 and 10 ppb.
- Canada (Health Canada, 2006, 118 drinks). Benzene was found above the Canadian drinking-water guideline of 5 micrograms per litre in four products, averaging 6.0–23.0 micrograms per litre. A follow-up survey in 2007 (139 drinks) found three products above 5 micrograms per litre.
- Korea (April 2006). The food and drug authority detected benzene in 27 of 30 (90%) vitamin-enriched drinks, ranging from 5.7 to 87.8 ppb, and advised a voluntary recall above 10 ppb.
The benchmarks to hold those numbers against: the US EPA maximum contaminant level (MCL) for benzene in drinking water is 5 ppb (5 micrograms per litre), and the FDA adopted the EPA MCL as the allowable level for bottled water. The WHO drinking-water guideline is 10 ppb. There is no benzene standard for beverages as such — the drinking-water numbers are the comparison everyone uses, not a legal limit for soft drinks.
A later retail check adds useful context. Nyman et al. (2010, Journal of Food Science) found that, except for cranberry juice, retail beverages contained no detectable benzene (under 0.05 nanograms per gram) at the end of shelf life. In model solutions, 24 hours of intense UVA raised benzene by up to 53% in non-UV-stabilised bottles, while UV-stabilised PET cut formation by about 13% relative to non-stabilised bottles; testing at 40 °C for 14 days best simulated end-of-shelf-life formation. The American Beverage Association published guidelines recommending accelerated testing (UV plus heat) to screen formulations.
The fair framing, then: benzene formation is real chemistry, demonstrated in the lab and found in a small minority of surveyed drinks in 2005–2007, mostly below drinking-water limits; the products found above the limit were reformulated; and regulators' position is that current levels do not pose a safety concern. The counter-point is also real: there is no ongoing public FDA product-testing programme cited, so the reassurance rests on surveys that are now twenty years old. One older dataset is sometimes quoted and should not be: the Total Diet Study's 1996–2001 figures — 19 of 24 diet-soda samples above 5 ppb, average 19 ppb, maximum 55 ppb — were generated with an analytical method that itself formed benzene during testing, so the FDA says those figures are not comparable.
Tip: if benzene is your concern, the practical move is narrow and works: choose drinks that do not list both sodium benzoate and ascorbic acid (vitamin C) together, and keep opened acidic drinks refrigerated and out of heat and sunlight.
Hyperactivity: the Southampton trial, and what it did not test
The second question about sodium benzoate comes from a single famous trial — and the most important fact about it is what it did not test.
McCann et al. (2007, The Lancet) — the Southampton study — was a randomised, double-blind, placebo-controlled trial in 153 three-year-olds and 144 eight- and nine-year-olds from the general population, not children with ADHD. After six weeks on a benzoate- and colour-free diet, the children drank, in random order across three weeks, two active mixes and a placebo drink, with washout weeks between. The challenge drinks were 300 mL a day for the three-year-olds and 625 mL a day for the older group, and the main outcome was a global hyperactivity aggregate combining observed behaviour with parent and teacher ratings, plus a computerised attention test for the older children.
The exact doses are the point of this section. Sodium benzoate was in both active mixes, at 45 mg a day, for both age groups. For three-year-olds, Mix A was 20 mg of colours (sunset yellow, carmoisine, tartrazine, ponceau 4R) plus 45 mg of sodium benzoate, and Mix B was 30 mg of colours plus 45 mg of sodium benzoate. For eight- and nine-year-olds, Mix A was 24.98 mg of colours plus 45 mg of sodium benzoate, and Mix B was 62.4 mg of colours plus 45 mg of sodium benzoate. The doses were framed as roughly the colouring in two 56-gram bags of sweets a day for the younger children, and two to four bags a day for the older ones.
The results were small and inconsistent. For three-year-olds, Mix A versus placebo gave a global hyperactivity effect size of 0.20 (95% CI 0.01–0.39, p=0.044), rising to 0.32 (0.05–0.60, p=0.02) in children who drank at least 85% of the juice with no missing data; Mix B was not significant in three-year-olds. For eight- and nine-year-olds who drank at least 85% with no missing data, Mix A gave 0.12 (0.02–0.23, p=0.023) and Mix B gave 0.17 (0.07–0.28, p=0.001). Effect sizes of 0.12–0.32 are small by any standard — and the two mixes did not behave the same way in the two age groups.
Here is the limitation, stated in the paper itself and by EFSA: the trial tested mixtures. There is no benzoate-only arm — the study cannot show that sodium benzoate alone affects behaviour. As the American Academy of Pediatrics' summary put it, "No distinction could be made between the effects of food coloring versus preservative on child behavior." The paper's own interpretation is carefully worded: "Artificial colours or a sodium benzoate preservative (or both) in the diet result in increased hyperactivity in 3-year-old and 8/9-year-old children in the general population." Colours, or benzoate, or both — the trial cannot say which.
EFSA's 2008 assessment of the study concluded it provided "limited evidence that the mixtures of additives tested had a small effect on the activity and attention of some children", but that the effects were not consistent across the two age groups and the two mixtures, and that given the uncertainties — lack of consistency, relative weakness of effect, no information on clinical significance — the findings "could not be used as a basis for altering the ADI of the respective food colours or sodium benzoate". There was also, EFSA noted, no plausible biological mechanism.
The regulatory consequence is precise, and it is often misreported. The EU response was a warning label on the six azo colours only (E102, E104, E110, E122, E124, E129) — "may have an adverse effect on activity and attention in children" — under Regulation (EC) No 1333/2008 Annex V. Sodium benzoate carries no such warning.
An earlier, smaller study points the same way. Bateman et al. (2004) gave 277 three-year-olds a drink containing 20 mg of artificial colourings plus 45 mg of sodium benzoate daily, against placebo, in a three-week crossover after a week of additive-free diet. Hyperactive behaviour fell during withdrawal and rose more during the active period than the placebo period by parental reports — but there were no significant differences on objective clinic testing.
Two common misreadings to avoid. First, the meta-analyses sometimes quoted in this debate are about colours, not benzoate: Nigg et al. (2012) estimated that 8% of children with ADHD may have symptoms related to synthetic food colours — a colours figure, not a benzoate figure, and it must not be presented as one. Second, "Southampton doses of 25 mg/kg/day" appears on a commercial allergy site; it is inconsistent with the paper, which used a fixed 45 mg a day of sodium benzoate for both age groups, and it is not used in this article. The trial tested colours and benzoate together — not flavours: no flavour-specific hyperactivity trial was found, as our artificial flavours guide explains.
The fair summary: one good trial found small, inconsistent effects of colour-plus-benzoate mixtures on hyperactivity scores; no study isolates sodium benzoate; EFSA judged the evidence too limited to change the intake limit; there is no established mechanism and no ADHD-causation claim. Parents of sensitive children may still reasonably choose to avoid the combination — that is a precautionary choice, not a proven effect.
Related: Artificial colours — the synthetic dyes behind the same hyperactivity debate, and the six colours that carry the EU warning label.
Allergy and intolerance: real, uncommon, and hard to prove
Reactions to benzoates exist, but they are uncommon, hard to prove, and usually not allergy in the sense most people mean the word.
The mechanism is a pseudo-allergy, not a true allergy. Benzoic acid directly degranulates mast cells — the cells that release histamine — without IgE antibodies, so standard allergy skin tests come back negative. True IgE-mediated allergy to benzoates is rare. Type IV contact dermatitis also occurs, but is less common than the urticaria mechanism, and benzoic acid was not a sensitiser at 2% in the studies reviewed by the Cosmetic Ingredient Review (CIR).
The best challenge data put the scale in perspective. In a double-blind placebo-controlled challenge study, 47 people with repeated acute urticaria or angio-oedema after meals containing sodium benzoate were challenged with the ingredient. Only 1 of 47 (2%) reacted to 75 mg of sodium benzoate without reacting to placebo. The authors' conclusion: the percentage of repeated urticaria or angio-oedema episodes induced by sodium benzoate is very low (2%), and other causes should be evaluated. Note what that 2% is: 2% of people who already have repeated unexplained urticaria episodes — not 2% of the general population. There is no verified general-population prevalence figure for benzoate intolerance.
The aspirin connection is the best-documented pattern. Rosenhall (1982) ran 1,868 oral challenge tests in 504 patients with asthma or rhinitis, and found that sodium benzoate and other benzoic compounds, plus tartrazine and other azo dyes, "were found to give similar but milder adverse reactions than acetylsalicylic acid". No patients reacted to potassium sorbate. Intolerance was most frequent in patients with a history of adverse reactions to analgesics or food — but also in 6% of patients investigated for other reasons. Patients with aspirin-exacerbated respiratory disease (Samter's triad — aspirin-sensitive asthma, roughly 10–20% of adult asthmatics) often show cross-reactivity to benzoates through shared pathways involving arachidonic-acid metabolism and COX-1 inhibition, so benzoate ingestion may trigger wheezing in these patients — though that prevalence figure comes from a commercial source and should be treated as approximate.
The Danish EPA's toxicological evaluation records the case-report end of the spectrum: in sensitive persons, oral doses of 10–250 mg of benzoic acid (under 4 mg/kg of body weight) have induced eczema, and anaphylaxis is documented only in isolated cases — a young woman after a meal containing sodium benzoate, with a provocation test of 160 mg inducing localised urticaria, and a worker who developed anaphylaxis after oral exposure to 425 mg (about 5 mg benzoic acid per kilo). Isolated case reports are the weakest evidence tier, but they are the reason the guidance for sensitive people is to take the possibility seriously rather than to dismiss it.
For cosmetics, the CIR's 2001 final report concluded that benzyl alcohol, benzoic acid and sodium benzoate are safe in cosmetic formulations at concentrations up to 5%, with a caution that manufacturers should consider the non-immunologic contact-urticaria phenomenon in formulations for infants and children. A 2017 CIR supplement concluded the ingredients "are safe in the present practices of use and concentration".
The honest verdict: intolerance exists but is uncommon and hard to prove — the best challenge data put confirmed benzoate-induced urticaria at about 2% of people who already have repeated unexplained urticaria episodes. Anaphylaxis is documented only in isolated case reports. The aspirin and asthma link is a recognised cross-reactivity pattern in challenge studies, not a proven population risk.
What regulators say
Every major authority permits sodium benzoate within limits — but the limits, and the intake numbers behind them, no longer agree with each other. That disagreement is the regulatory story of this ingredient.
United States. Sodium benzoate is generally recognised as safe under 21 CFR 184.1733, affirmed in 1977 and amended in 1984, used as an antimicrobial agent and as a flavouring agent and adjuvant. The regulation says the ingredient "is used in food at levels not to exceed good manufacturing practice. Current usage results in a maximum level of 0.1 percent in food." Read that carefully: the 0.1% is a statement of what manufacturers currently use, not a hard legal cap — the regulation sets no numeric limit, and the FDA adds that it "has not determined whether significantly different conditions of use would be GRAS". In cosmetics there is no FDA numeric limit; the CIR position is safe up to 5% (2001) and "safe in the present practices of use and concentration" (2017). No FDA warning, ban or reassessment of sodium benzoate was found.
The intake-limit split. EFSA's 2016 re-evaluation set a group acceptable daily intake (ADI) of 5 mg/kg of body weight per day, expressed as benzoic acid, for benzoic acid and its sodium, potassium and calcium salts — derived from a four-generation reproductive-toxicity study in rats with a no-observed-adverse-effect level (NOAEL) of 500 mg of benzoic acid per kilo per day, using an uncertainty factor of 100. EFSA found short-term and subchronic toxicity low, no genotoxicity concern, and no indication of carcinogenic potential. In 2021, the WHO's Joint Expert Committee on Food Additives (JECFA) moved the other way: at its 92nd meeting it withdrew the previous group ADI of 0–5 mg/kg and set a group ADI of 0–20 mg/kg per day, expressed as benzoic-acid equivalents, using a chemical-specific adjustment factor of 2 for interspecies toxicokinetic variation instead of the default 4.0. The old 0–5 limit dated from 1983 and had been maintained in 1996 and 2001. EFSA (5) and JECFA (0–20) now differ by a factor of four — and any source still quoting "JECFA ADI 0–5" as current is out of date.
What those numbers mean in practice: at EFSA's 5 mg/kg a day, a 20 kg child has an ADI of 100 mg a day as benzoic acid (about 118 mg a day as sodium benzoate), and a 60 kg adult has 300 mg a day as benzoic acid (about 354 mg as sodium benzoate). And EFSA's own 2016 exposure work is the reason the question is not closed for children: the group ADI was exceeded in the brand-loyal scenario, in particular for toddlers and children consuming flavoured drinks on a regular basis, and counting additional exposure from carry-over, intake could rise two- to three-fold for all high-level consumers, producing ADI exceedance in toddlers and children even in the non-brand-loyal scenario. The main contributing categories were unprocessed fruits and vegetables and flavoured drinks. That is an exposure estimate, not a finding of harm — but it is a regulator's own finding, and no EU limit was lowered as a result of that opinion. For context, JECFA's 2015 exposure check under the old ADI found mean exposure for soft-beverage consumers of 0.3–4.1 mg/kg a day for toddlers and young children, 0.2–2.7 for other children and adolescents, and 0.1–1.7 for adults — none exceeding the old upper bound. The WHO's older CICAD 26 estimates put national mean benzoate intakes at 0.18 mg/kg a day (Japan) to 2.3 (USA), with high consumers at 7.3 (USA) and 14 (China) — both above the then-ADI of 0–5. Codex's General Standard for Food Additives provisions for benzoates include 300 mg/kg in dairy-based desserts, 1,000 mg/kg in fat spreads and dried vegetables, and 2,000 mg/kg in vegetables.
European Union — food. Sodium benzoate is authorised as E211 under Regulation (EC) No 1333/2008 Annex II, in 44 food categories, at 150–5,000 mg/kg. The range matters, and so does a trap inside it: most limits cover sorbates and benzoates combined, expressed as the free acid — so a quoted limit is usually the combined sorbate-plus-benzoate ceiling, not a benzoate-only ceiling. The headline figures: flavoured drinks 150 mg/l (benzoates only); fruit and vegetables in vinegar, oil or brine 2,000 mg/kg (combined); sauces 1,000–2,000 mg/kg (combined); semi-preserved fish including roe 2,000 mg/kg (combined); and liquid egg 5,000 mg/kg (combined) — the highest EU limit.
| Food category | EU limit | What the limit covers |
|---|---|---|
| Flavoured drinks | 150 mg/l | Benzoates only |
| Fruit and vegetables in vinegar, oil or brine | 2,000 mg/kg | Sorbates + benzoates combined |
| Sauces | 1,000–2,000 mg/kg | Sorbates + benzoates combined |
| Semi-preserved fish, including roe | 2,000 mg/kg | Sorbates + benzoates combined |
| Liquid egg | 5,000 mg/kg | Sorbates + benzoates combined — the highest EU limit |
European Union — cosmetics. Regulation (EC) No 1223/2009 Annex V, entry 1, allows benzoic acid and its sodium salt at 0.5% (as acid) in leave-on products, 1.7% (as acid) in oral products, and 2.5% (as acid) in rinse-off products. The percentages are expressed as benzoic acid, not as the sodium salt — 0.5% as acid is about 0.59% as sodium benzoate.
Australia and New Zealand. Benzoic acid and its sodium, potassium and calcium salts are permitted under FSANZ Schedule 15. The maximum permitted levels include 400 mg/kg in flavoured drinks and fruit juices, 1,000 mg/kg in sauces and dressings, 2,500 mg/kg in semi-preserved fish, and 9,000 mg/kg in rennetting-enzyme preparations — the highest FSANZ level. Benzoates do not appear in the permitted-additive list for infant formula products (Schedule 15 item 13.1). On labels, sodium benzoate appears as "preservative (211)" or "sodium benzoate (211)".
Canada. Sodium benzoate is a permitted preservative under Health Canada's List of Permitted Preservatives (List 11), item S.3, for jams, jellies, marmalades, fruit juices, pickles, relishes, tomato products, margarine, marinated fish and meat products, unstandardised foods and oyster-flavoured sauces. The limit is 1,000 ppm, calculated as benzoic acid, whether used singly or combined with benzoic acid or potassium benzoate; for flavour essences and extracts the limit is good manufacturing practice.
US state law — benzoates are not banned anywhere. California's AB 418 (the California Food Safety Act, effective 1 January 2027) bans exactly four substances — brominated vegetable oil, potassium bromate, propylparaben and Red dye 3. West Virginia's HB 2354, which deems foods containing BHA, propylparaben and seven FD&C colours "poisonous or injurious", had its statewide enforcement preliminarily enjoined on 23 December 2025 as unconstitutionally vague; its school-meals provision is unaffected. West Virginia's HB 4852 (2026) lists the same additives and had advanced to the full Senate, not enacted. New York's S1239, which bans Red 3, potassium bromate and propylparaben, had passed both houses and was awaiting governor action as of early November 2026. No enacted US state law and no pending state bill found in this research names sodium benzoate or benzoates — the state bans target colours, BHA and propylparaben, as our BHA & BHT guide explains. California's stalled AB 2034 would have deemed ingredients unsafe without a recognised safe harbour, including recognition in federal regulations as GRAS — a harbour sodium benzoate would meet via 21 CFR 184.1733.
Structural change, not substance-specific. The FDA's proposed rule of 10 August 2026 (91 FR 51834, comments due 9 December 2026) would convert the voluntary GRAS notification programme into a mandatory one. It revokes no GRAS substance and does not name sodium benzoate — and because sodium benzoate's status rests on a codified Part 184 affirmation rather than a self-affirmed determination, it is among the least affected substances.
Related: Maltodextrin — another codified-GRAS ingredient, and the same 2026 GRAS-reform proposal explained in full.
The debate
The case that the concern is overstated
Every regulator that has examined sodium benzoate has left it permitted — and in 2021 the WHO's expert committee moved the intake limit up, not down, quadrupling it from 0–5 to 0–20 mg per kilo a day on a revised toxicokinetic factor. EFSA's own 2016 opinion found low toxicity, no genotoxicity concern and no carcinogenic potential. The benzene chemistry is real, but the surveys found only a small minority of drinks above drinking-water limits, and those products were reformulated to under 1.5 ppb; the FDA's conclusion was that the levels found do not pose a safety concern. The Southampton trial could not isolate benzoate from the colours it was mixed with, and EFSA called the evidence "limited" — not a basis for changing the intake limit. Allergy is confirmed in about 2% of people who already have repeated unexplained urticaria, not in the general population. Even the advocacy group EWG rates the cancer, allergy and developmental concerns LOW on its own Skin Deep page. The ingredient is approved everywhere, at doses far above what most people consume.
The case that the questions are still live
Nobody disputes the chemistry: benzene is a Group 1 human carcinogen, and it can form from two lawful additives sitting in the same bottle — the 1993 paper is not contested, only its real-world magnitude. EFSA's own 2016 exposure work shows toddlers and children who regularly drink flavoured drinks can exceed the intake limit, and no EU limit was lowered as a result of that opinion — Europe's exposure work points one way while the WHO committee's toxicology review points the other, and the fourfold gap between them is not a rounding error. The Southampton trial did find small but real behavioural effects, and "limited evidence" is not the same as "no effect" — the mixtures were never unpicked, so benzoate has never actually been cleared on its own. A small minority — aspirin-sensitive people, people with chronic urticaria — genuinely react, and the reassurance about benzene rests on surveys that are now twenty years old, with no ongoing public testing programme cited. Approved everywhere is true; settled is not.
Who should avoid it
Sodium benzoate is not an ingredient most people need to avoid. The people who should think about it are specific, and the reasons are different for each of them.
- Parents of toddlers and young children who drink a lot of flavoured drinks. This is the one exposure flag that comes from a regulator's own work: EFSA's 2016 assessment found the ADI could be exceeded by toddlers and children who regularly drink flavoured drinks, and by high consumers generally once carry-over is counted. Water and milk are the straightforward swaps for everyday thirst.
- People with aspirin sensitivity or aspirin-exacerbated respiratory disease. Challenge studies show benzoates can give similar but milder reactions than aspirin, and cross-reactivity runs through shared arachidonic-acid and COX-1 pathways. If aspirin triggers your asthma, discuss benzoate-containing foods and drinks with your clinician rather than self-testing.
- People with chronic, unexplained urticaria. The best challenge data put confirmed benzoate-induced urticaria at about 2% of people who already have repeated unexplained episodes — low, but not zero, and worth establishing by proper challenge testing rather than by guessing from a food diary.
- People who want to avoid the benzene question entirely. Choose drinks that do not list both sodium benzoate and ascorbic acid (vitamin C) together, keep opened acidic drinks refrigerated, and do not store them hot or in sunlight.
- People with non-immunologic contact urticaria from cosmetics. Benzoic acid can cause wheals, redness and itching within 30–45 minutes of skin application in sensitive people. The CIR cautions manufacturers about this phenomenon in formulations for infants and children, and the EU's leave-on limit (0.5% as acid) is the strictest of its three cosmetic caps for a reason.
For everyone else, the numbers are reassuring in a way that is worth stating plainly: at EFSA's limit, a 60 kg adult would need about 300 mg of benzoic acid a day — roughly 354 mg as sodium benzoate — to reach the ADI, and typical intakes sit well below it. The argument for avoiding benzoate is a targeted one, for targeted people, not a general one.
How to spot it on a label
Sodium benzoate is one of the easier ingredients to spot, because it is declared under its own name. The hard part is knowing what the name does and does not tell you.
- The name itself. Look for "sodium benzoate" in the ingredients list, in descending order of weight — if it is near the top, there is a lot of it. In the EU it may appear as "E211"; in Australia and New Zealand as "preservative (211)" or "sodium benzoate (211)". It has no hidden synonym.
- The vitamin C line. If benzene is your concern, the second line to check is ascorbic acid (vitamin C). A drink listing both benzoate and ascorbic acid is the combination the 1993 chemistry describes; a drink listing benzoate without vitamin C is not.
- The numbers that differ by country. The same flavoured drink can lawfully contain up to 150 mg/l of benzoates in the EU, 400 mg/kg in Australia and New Zealand, and about 0.1% (around 1,000 mg/kg) under current US use. The limits are not the same because the regulators are not the same — not because the ingredient changes at the border.
- What the label will not tell you. Benzoate does not appear on the nutrition panel — only in the ingredients list — and there is no way to tell from a label how much benzene, if any, has formed in the bottle. Storage matters more than the label here: heat and light are the drivers, and they leave no trace on the packaging.
- In cosmetics. Check leave-on products first if you are sensitive: the EU allows 0.5% (as acid) in leave-on products against 2.5% in rinse-off products, and the lower cap is the one that reflects skin contact time.
Alternatives: what replaces it
There is no drop-in "healthy sodium benzoate", because the job it does — cheap, effective preservation of acidic foods — is the job itself. The honest swaps are formulation options, and the list below is general trade knowledge, not a ranked recommendation.
- Potassium sorbate — the closest direct swap. It inhibits moulds, yeasts and some bacteria, works in acidic foods and beverages, and is effective to roughly pH 6.0–6.5, wider than benzoates' pH below 4.5. It does not form benzene with vitamin C. Notably, in Rosenhall's 1982 challenge study, no patients reacted to potassium sorbate.
- Sorbic acid — the parent acid of potassium sorbate, used in cheese, yogurt, dried meats and beverages.
- Calcium propionate — a mould and bacteria inhibitor, mainly for baked goods, tortillas and cheese.
- Lactic acid and vinegar (acetic acid) — fermentation-derived acids used in pickles, sauerkraut, yogurt and salad dressings.
- Nisin — a naturally occurring bacteriocin from Lactococcus lactis, active against Gram-positive bacteria, approved in more than 50 countries and used in dairy, canned foods and meat alternatives.
- Natamycin — an antifungal from Streptomyces natalensis, suited to dairy and cheese surfaces.
- Rosemary extract — a plant-derived antimicrobial and antioxidant, mainly for oils, meats and snacks, and EU-authorised as E392.
- High-pressure processing (HPP) — a non-thermal preservation technique that needs no additive at all.
The trade-offs are the reason benzoate survives. Natural options tend to be less potent and more expensive than benzoate or sorbate, and no single alternative covers benzoate's full spectrum — yeast plus mould plus some bacteria at low pH and low cost — so manufacturers usually combine approaches. A "preservative-free" label often means a different preservation system (heat, pressure, acidity, refrigeration), not no preservation at all.
For the shopper, the practical swaps are simpler than the formulation list: choose drinks without both benzoate and vitamin C on the label if benzene is the concern; refrigerate opened acidic drinks and do not store them hot or in sunlight; pick fresh and whole foods over preserved ones; and for home preserving, vinegar, lemon juice, salt and refrigeration are the traditional routes — with no efficacy figures claimed for them here.
Frequently asked questions
Is sodium benzoate safe?
Every major regulator permits it within limits: it is generally recognised as safe in the US, authorised as E211 in the EU, and permitted in Australia, New Zealand and Canada. EFSA's intake limit is 5 mg per kilo of body weight a day, and the WHO committee raised its own limit to 20 mg/kg in 2021. For most people at normal intakes it sits well inside both limits; the exposure flag is for toddlers and children who drink a lot of flavoured drinks.
Can sodium benzoate form benzene?
Yes — when a drink also contains vitamin C, with heat, light, storage time and trace metals increasing formation. The chemistry was shown in 1993, and the 2005–2007 surveys found ten of about 200 US products above the 5 ppb drinking-water limit; those products were reformulated to under 1.5 ppb. Regulators concluded current levels do not pose a safety concern. There is no benzene standard for beverages as such — the 5 ppb figure is the drinking-water limit.
Does sodium benzoate cause hyperactivity or ADHD?
No study shows that. The 2007 Southampton trial found small effects from mixtures of colours plus sodium benzoate (45 mg a day in both mixes), but it had no benzoate-only arm, so it cannot show benzoate alone affects behaviour. EFSA judged the evidence too limited to change the intake limit, and the EU warning label that followed applies to the six azo colours only — not to benzoate. There is no ADHD-causation claim.
Is sodium benzoate banned anywhere?
No. California's AB 418 bans four substances (brominated vegetable oil, potassium bromate, propylparaben, Red 3); West Virginia's HB 2354 and HB 4852 and New York's S1239 target colours, BHA and propylparaben. No enacted US state law and no pending state bill found in this research names sodium benzoate or benzoates.
What is E211?
E211 is sodium benzoate's number in the EU's food-additive system — the same ingredient as 'sodium benzoate' on a US label and '211' on an Australian or New Zealand label. In the EU it is authorised in 44 food categories at 150–5,000 mg/kg, with most limits covering sorbates and benzoates combined, expressed as the free acid.
Is sodium benzoate safe in cosmetics?
The EU allows 0.5% (as benzoic acid) in leave-on products, 1.7% in oral products and 2.5% in rinse-off products. The Cosmetic Ingredient Review concluded the benzoate ingredients are safe in cosmetics up to 5%, and in 2017 that they are safe in present practices of use — with a caution about non-immunologic contact urticaria in formulations for infants and children. Sensitive people can develop wheals, redness and itching where it is applied.
Checked against 21 CFR 184.1733, EFSA Journal 2016;14(3):4433, the 92nd JECFA (2021), Regulation (EC) No 1333/2008 and FSANZ Schedule 15 · Last updated 2026-11-13 · How we assess ingredients
Sources (24)
- US Food and Drug Administration — 21 CFR 184.1733, Sodium benzoate (eCFR via Cornell LII): the GRAS definition, antimicrobial and flavouring uses, and the current-usage statement of a maximum level of 0.1 percent in food under good manufacturing practice
- US Food and Drug Administration — Questions and Answers on the Occurrence of Benzene in Soft Drinks and Other Beverages (19 May 2006, updated 27 Mar 2007 and 12 Jul 2007): CFSAN survey of almost 200 samples, ten products above 5 ppb, reformulated products under 1.5 ppb, and the conclusion that the levels found do not pose a safety concern
- EFSA ANS Panel (2016), Re-evaluation of benzoic acid and sodium benzoate as food additives, EFSA Journal 14(3):4433: group ADI 5 mg/kg bw/day as benzoic acid (NOAEL 500 mg/kg bw/day, uncertainty factor 100); no genotoxicity concern; no carcinogenic potential; ADI exceeded for toddlers and children in the brand-loyal scenario (flavoured drinks)
- World Health Organization — JECFA database, benzoic acid (evaluation year 2021): group ADI 0–20 mg/kg bw withdrawn and re-established at the 92nd JECFA, withdrawing the previous 0–5 group ADI; sodium benzoate INS 211, CAS 532-32-1
- FAO — 92nd JECFA Chemical and Technical Assessment (2021): 'A group ADI 0-5 mg/kg bw for benzoic acid and its salts was established at the 27th JECFA (1983). The ADI was withdrawn and re-established to 0-20 mg/kg… at the 92nd JECFA (2021)'
- German Federal Institute for Risk Assessment (BfR), Expert Opinion No. 013/2006 (1 Dec 2005): summary of Gardner & Lawrence 1993 — benzene maximal at pH 2 (about 37 nmol/L, roughly 2.9 µg/L), lower at higher pH; 'Whether and to what extent benzene is actually formed in the corresponding foods cannot… be reliably assessed on the basis of the data available'; flavoured drinks limit 150 mg/l
- McCann D et al. (2007), Food additives and hyperactive behaviour in 3-year-old and 8/9-year-old children in the community, The Lancet 370(9598):1560–1567: 153 three-year-olds and 144 eight/nine-year-olds; sodium benzoate 45 mg/day in both active mixes; Mix A effect size 0.20 in three-year-olds, 0.12–0.17 in eight/nine-year-olds; no benzoate-only arm
- EFSA (2008), EFSA evaluates Southampton study on food additives and child behaviour: the AFC Panel found 'limited evidence' of a small effect of the tested mixtures and concluded the findings 'could not be used as a basis for altering the ADI' of the food colours or sodium benzoate
- Nigg JT et al. (2012), Meta-analysis of attention-deficit/hyperactivity disorder or attention-deficit/hyperactivity disorder symptoms, restriction diet, and synthetic food color additives, Journal of the American Academy of Child & Adolescent Psychiatry 51(1):86–97.e8: synthetic food colours only — parent reports g=0.18, 'an estimated 8% of children with ADHD may have symptoms related to synthetic food colors'; not a benzoate figure
- Danish Environmental Protection Agency (1999), Toxicological evaluation of benzoic acid and sodium benzoate: conversion factor 100 mg sodium benzoate = 84.7 mg benzoic acid; hypersensitive persons reacting to 10–250 mg benzoic acid; anaphylaxis case reports (Michils 1991, Pevny 1981)
- Rosenhall L (1982), Evaluation of intolerance to analgesics, preservatives and food colorants, European Journal of Respiratory Diseases 63:410–419 (PubMed): 504 patients with asthma or rhinitis, 1,868 oral challenge tests; benzoate reactions similar but milder than aspirin; no patients reacted to potassium sorbate
- Nair B (2001), Final report on the safety assessment of benzyl alcohol, benzoic acid, and sodium benzoate, International Journal of Toxicology 20 (suppl 3):23–50 (CIR): safe in cosmetic formulations up to 5%, with a caution about non-immunologic contact urticaria in infant and child formulations
- European Union — Regulation (EC) No 1333/2008, Annex II (Union list of authorised food additives): E 210–213 benzoic acid — benzoates, including E211 sodium benzoate; limits mostly combined sorbates + benzoates, expressed as the free acid
- European Union — Regulation (EC) No 1223/2009, Annex V (consolidated text, 16 Apr 2015): entry 1, benzoic acid and its sodium salt — leave-on 0.5% (acid), oral 1.7% (acid), rinse-off 2.5% (acid)
- Food Standards Australia New Zealand — Food Standards Code, Schedule 15 (Substances that may be used as food additives): benzoates 210–213 permitted; flavoured drinks 400 mg/kg, sauces 1,000 mg/kg, semi-preserved fish 2,500 mg/kg, rennetting enzymes 9,000 mg/kg; no benzoate entry in the infant-formula list (item 13.1)
- California Legislative Information — AB 418 (2023–2024), California Food Safety Act: bans exactly four substances (brominated vegetable oil, potassium bromate, propylparaben, Red 3); sodium benzoate not named
- National Agricultural Law Center — Preliminary Injunction Halts Enforcement of West Virginia's Food Dye Ban (23 Dec 2025): HB 2354 statewide enforcement preliminarily enjoined (International Association of Color Manufacturers v. Singh); school-meals provision unaffected
- World Health Organization — Concise International Chemical Assessment Document 26, Benzoic Acid and Sodium Benzoate (2000): national mean intakes 0.18 mg/kg bw/day (Japan) to 2.3 (USA); high consumers 7.3 (USA) and 14 (China) mg/kg bw/day
- Nyman PJ et al. (2010), Evaluation of Accelerated UV and Thermal Testing for Benzene Formation in Beverages Containing Benzoate and Ascorbic Acid, Journal of Food Science: retail beverages no detectable benzene (under 0.05 ng/g) at end of shelf life except cranberry juice; 24 h intense UVA raised benzene up to 53% in model solutions in non-UV-stabilised bottles
- Health Canada — Benzene in Soft Drinks and Other Beverages: follow-up survey (2007): three of 139 drinks above the 5 µg/L drinking-water guideline; 2006 survey found four products averaging 6.0–23.0 µg/L
- EWG Food Scores — Sodium Benzoate (ADVOCACY source): 'Overall Sodium Benzoate is of moderate concern in food', citing an SCF 2002 developmental-toxicity finding in animals at high doses and an in-vitro genotoxicity study (Zengin et al. 2011)
- EWG Skin Deep — Sodium Benzoate (ADVOCACY source): concerns LOW for cancer, allergies/immunotoxicity and developmental/reproductive toxicity, MODERATE for use restrictions; data availability 'Fair'
- Healthline — Sodium Benzoate: Uses, Dangers, and Safety (secondary overview, RD-reviewed): reports that EWG scores the additive 3 on a 0–10 scale, and describes inflammation, oxidative stress, obesity, ADHD and allergy links as preliminary
- Food Additive Index — Sodium Benzoate (secondary compilation of Regulation (EC) No 1333/2008 Annex II, FSANZ Schedule 15 and Health Canada List 11 tables): EU category limits, FSANZ maximum permitted levels, and the FDA Substances Added to Food inventory entry
