The dose makes the poison — but which doses, exactly?

Every food preservative has an Acceptable Daily Intake (ADI) established by regulatory bodies like the European Food Safety Authority (EFSA) and the Joint FAO/WHO Expert Committee on Food Additives (JECFA). The ADI represents the amount of a substance that can be consumed daily over a lifetime without appreciable health risk, expressed in milligrams per kilogram of body weight per day (mg/kg bw/day).

The ADI is not a cliff edge. Exceeding it on a single day does not cause harm. It is a chronic exposure threshold calculated by taking the No Observed Adverse Effect Level (NOAEL) from animal studies and dividing by a safety factor — typically 100 (10x for species extrapolation, 10x for individual human variation). This means the ADI already has a hundredfold safety margin built in.

What matters is your habitual intake pattern. A 70 kg adult with an ADI of 5 mg/kg bw/day can consume 350 mg daily for their entire life with no expected adverse effect. The question for each preservative is: does typical dietary exposure approach that threshold?

Preservative safety tier table

This ranking is based on weight of evidence from EFSA re-evaluations (2015-2025), FDA GRAS determinations, JECFA assessments, and published systematic reviews. “Safety tier” reflects the consensus position across major regulatory bodies, not any single study.

TierPreservativeE-NumberCAS NumberADI (mg/kg bw/day)Mechanism of ActionEvidence Quality
1 — Well-Established SafeCitric acidE33077-92-9Not specifiedpH reduction; chelation of metal ionsHigh
1Ascorbic acid (Vitamin C)E30050-81-7Not specifiedOxygen scavenging; free radical terminationHigh
1Tocopherols (Vitamin E)E306-E30959-02-92.0 (synthetic dl-alpha)Lipid peroxidation chain terminationHigh
1Acetic acidE26064-19-7Not specifiedpH reduction below microbial growth thresholdHigh
1Lactic acidE27050-21-5Not specifiedpH reduction; membrane disruptionHigh
1Rosemary extractE39220283-92-5 (carnosic acid)Not specified (EU); GRAS (US)Phenolic antioxidant; radical scavengingModerate
2 — Limited ConcernSodium benzoateE211532-32-15.0Inhibits anaerobic fermentation enzymesHigh
2Potassium sorbateE20224634-61-525.0Inhibits dehydrogenase enzymes in molds/yeastsHigh
2Calcium propionateE2824075-81-4Not specifiedInhibits mold enzyme systemsModerate
2Sulfur dioxide / SulfitesE220-E2287446-09-50.7SO2 disrupts disulfide bonds; enzyme inhibitionHigh
2NisinE2341414-45-51.0Pore formation in Gram-positive cell membranesModerate
2NatamycinE2357681-93-80.3Binds ergosterol in fungal membranesModerate
3 — Actively DebatedSodium nitriteE2507632-00-00.07NO release inhibits C. botulinum; cured meat color fixationHigh (contested)
3Sodium nitrateE2517631-99-43.7Converts to nitrite in vivo; same antimicrobial pathwayHigh (contested)
3Propyl gallateE310121-79-90.5Phenolic radical scavenging in fatsLow
4 — RestrictedBHA (butylated hydroxyanisole)E32025013-16-51.0 (EFSA) / 0.5 (JECFA)Phenolic chain-breaking antioxidant in lipidsModerate (concerns)
4BHT (butylated hydroxytoluene)E321128-37-00.25 (EFSA) / 0.3 (JECFA)Phenolic chain-breaking antioxidant in lipidsModerate (concerns)
4TBHQ (tert-butylhydroquinone)1948-33-00.7Donates hydrogen to lipid peroxyl radicalsLow

Preservative effectiveness by food type

Not every preservative works in every food matrix. pH, water activity, fat content, and target organisms determine which preservative is appropriate:

Food TypePrimary Spoilage ThreatMost Effective PreservativesIneffective ChoicesTypical Max Concentration
Acidic beverages (pH <4.5)Yeasts, moldsSodium benzoate (E211), potassium sorbate (E202)Nitrites (wrong target)150-250 mg/kg
Neutral beverages (pH >4.5)Bacteria, yeastsNisin (E234), sulfites (E220)Benzoate (inactive above pH 4.5)200-500 mg/kg
Bread and bakeryMolds (Rhizopus, Penicillium)Calcium propionate (E282), sorbic acid (E200)BHA/BHT (antioxidants, not antimicrobials)2000-3000 mg/kg
Cured meatsC. botulinum, ListeriaSodium nitrite (E250), nisin (E234)Sorbate (insufficient spectrum)80-150 mg/kg (nitrite)
CheeseMolds, surface bacteriaNatamycin (E235), sorbic acid (E200)Benzoate (pH too high)1-5 mg/dm2 (surface)
Cooking oils and fatsLipid oxidation (rancidity)BHA (E320), BHT (E321), TBHQ, tocopherols (E306)Benzoate (not an antioxidant)100-200 mg/kg
Dried fruitBrowning, moldsSulfur dioxide (E220)Nitrites (wrong application)500-2000 mg/kg
WineOxidation, wild yeasts, AcetobacterSulfites (E220-E228), sorbic acid (E200)Propionate (wrong matrix)150-400 mg/L (total SO2)
Fermented vegetablesUndesirable bacteria during fermentationLactic acid (E270), acetic acid (E260)Synthetic preservatives (inhibit desired fermentation)Self-generated

The nitrite question — nuance that headlines miss

Sodium nitrite (E250) is the most polarizing preservative in food science. The IARC classification of processed meat as Group 1 carcinogenic is frequently attributed to nitrite, but the mechanism is more specific: nitrite reacts with secondary amines under high heat to form N-nitroso compounds (NOCs), which are the actual carcinogens.

Three facts complicate the simple “nitrite is dangerous” narrative:

First, approximately 80% of dietary nitrate (which converts to nitrite) comes from vegetables — particularly beetroot, spinach, celery, and lettuce. A serving of spinach contains more nitrate than 467 hot dogs. The vegetable matrix appears protective because co-occurring vitamin C and polyphenols inhibit nitrosamine formation.

Second, nitrite serves a critical food safety function. It prevents Clostridium botulinum growth — the botulism organism. Removing nitrite from cured meats without an equivalent antimicrobial replacement increases the risk of a foodborne illness with a 5-10% fatality rate.

Third, the “uncured” or “no nitrite added” labels on products like natural hot dogs are misleading. These products use celery powder or cherry powder as nitrite sources. Celery-derived nitrite is chemically identical to synthetic sodium nitrite. Studies show “naturally cured” products sometimes contain higher nitrite levels than conventionally cured equivalents.

The evidence-based position: the concern is real but specific to high-temperature cooking of processed meat (grilling, frying), where nitrosamine formation peaks. Boiling or steaming cured meats produces significantly fewer NOCs.

Natural vs synthetic preservatives — honest comparison

The “natural preservative” market is projected at USD 3.1 billion by 2027. The marketing implies natural preservatives are safer. The evidence does not consistently support this assumption.

AttributeNatural Preservatives (rosemary, citrus extract, vinegar, nisin)Synthetic Preservatives (BHA, BHT, sodium benzoate, sorbate)
Regulatory safety dataLess extensive (shorter history of formal evaluation)More extensive (decades of ADI studies, re-evaluations)
Antimicrobial spectrumGenerally narrower; organism-specificBroader; effective against multiple spoilage types
Concentration requiredTypically 2-10x higher for equivalent effectLower effective concentrations
Consistency batch-to-batchVariable (depends on plant source, extraction method)Highly consistent (synthetic purity >99%)
Flavor impactOften significant (rosemary taste, vinegar acidity)Minimal at effective concentrations
Cost3-15x more expensive per kgLow cost, commodity chemicals
Shelf life extensionModerate (days to weeks shorter than synthetic equivalents)Superior (months of additional shelf life)
Consumer perceptionPositive (“clean label”)Negative (despite equivalent or better safety data)

The uncomfortable truth: calcium propionate (E282) — a synthetic preservative — is chemically identical to propionic acid produced naturally by Propionibacterium during Swiss cheese aging. Lactic acid (E270) sold as a preservative is produced by bacterial fermentation — the same process that generates it in yogurt. The natural/synthetic distinction is a marketing category, not a safety category.

Where natural preservatives genuinely excel: rosemary extract (E392) and tocopherols (E306-E309) are effective lipid antioxidants with strong safety profiles. They are replacing BHA and BHT in many reformulated products not because they are “natural” but because they work comparably and carry no IARC classification.

EU vs US vs Singapore — regulatory status comparison

PreservativeE-NumberEU (EFSA)US (FDA)Singapore (SFA)Maximum Permitted Level (EU)
Sodium benzoateE211AuthorizedGRASPermitted150-2000 mg/kg (food dependent)
Potassium sorbateE202AuthorizedGRASPermitted300-2000 mg/kg
Sodium nitriteE250Authorized (restricted)Approved (restricted)Permitted (restricted)150 mg/kg (ingoing)
SulfitesE220-E228Authorized (labeling >10 mg/kg)GRAS (labeling >10 ppm)Permitted (labeling required)50-2000 mg/kg
BHAE320AuthorizedGRASPermitted200 mg/kg fat
BHTE321Authorized (lower ADI)GRASPermitted100 mg/kg fat
TBHQAuthorized (limited use)ApprovedPermitted200 mg/kg fat
NisinE234AuthorizedGRASPermitted3-12.5 mg/kg
NatamycinE235Authorized (surface treatment only)Approved (cheese)Permitted1 mg/dm2 surface
Propyl gallateE310Authorized (under review)GRASPermitted200 mg/kg fat
Calcium propionateE282AuthorizedGRASPermitted3000 mg/kg (bread)
Rosemary extractE392AuthorizedGRASPermitted100-400 mg/kg (as carnosic acid)

The EU and US share a fundamental structural difference: the EU requires pre-market authorization and periodic re-evaluation (Regulation EC 1333/2008). The FDA allows manufacturer self-determination of GRAS status without mandatory notification. BHA carries a GRAS designation from 1958 that has never been formally revoked, despite IARC Group 2B classification in 1986.

Singapore’s SFA (Singapore Food Agency) generally aligns with Codex Alimentarius standards, accepting both EU and US safety assessments. The Food Regulations (Cap. 283, Rg 1) specify permitted preservatives in Schedule 7, with maximum levels aligned to Codex General Standard for Food Additives (GSFA). Singapore is notable for stricter enforcement of labeling requirements than many ASEAN neighbors.

Practical exposure analysis — can you reach the ADI?

For a 70 kg adult, here is how much of a single food you would need to consume daily to reach the ADI:

PreservativeADI (70 kg adult total)Typical ConcentrationFood Amount to Reach ADIRealistic Daily Intake% of ADI at Typical Intake
Sodium benzoate350 mg150-250 mg/kg in soft drinks1.4-2.3 L soft drink330 mL (one can)14-24%
Potassium sorbate1750 mg1000-2000 mg/kg in cheese875 g-1.75 kg cheese40 g (one slice)2-5%
Sodium nitrite4.9 mg80-150 mg/kg in cured meat33-61 g cured meat50 g (2 slices ham)82-152%
BHT17.5 mg100-200 mg/kg in cereal88-175 g cereal40 g (one bowl)23-46%
Sulfites49 mg100-300 mg/L in wine163-490 mL wine150 mL (one glass)31-92%
BHA70 mg100-200 mg/kg in chewing gum350-700 g gum3 g (one piece)<1%
Calcium propionateNot specified2000-3000 mg/kg in bread80 g (2 slices)

Sodium nitrite stands out: its ADI can be reached or exceeded with a single serving of cured meat. Sulfites approach the ADI with moderate wine consumption. All other common preservatives require quantities far beyond normal eating patterns. This is consistent with the epidemiological data identifying processed meat and sulfite-containing beverages as specific concerns rather than preserved food in general.

Honest limitations of this analysis

Several factors complicate clean ranking of preservative safety:

Cumulative exposure is poorly studied. ADIs are set for individual compounds. A person consuming sodium benzoate in a soft drink, potassium sorbate in bread, sulfites in wine, and nitrites in lunch meat simultaneously has no combined ADI to reference. Additive toxicology — the study of mixture effects — remains underdeveloped.

The benzene formation question. Sodium benzoate (E211) can react with ascorbic acid (E300) in acidic beverages to form benzene, a known carcinogen (IARC Group 1). Levels detected in soft drinks (1-20 ppb) are below the WHO drinking water guideline of 10 ppb in most but not all samples. The FDA tested 200 soft drinks in 2006 and found 5 exceeded the 5 ppb US drinking water standard. Industry reformulation has reduced but not eliminated this issue.

Evidence gaps in newer preservatives. Natamycin (E235) and nisin (E234) have shorter regulatory histories than benzoate or sorbate. Their safety profiles look favorable but are based on fewer long-term studies. The precautionary response is not to avoid them but to note the difference in evidence depth.

Individual variation. Sulfite sensitivity affects 5-10% of asthmatic individuals. This is not captured by the population-level ADI. If you have asthma and experience bronchospasm after wine, dried fruit, or restaurant salads (sulfite-sprayed to maintain freshness), your personal threshold is lower than the ADI suggests.

The evidence does not support blanket preservative avoidance. Tier 1 and Tier 2 preservatives have safety margins so large that typical dietary exposure is a small fraction of the ADI. The actionable concerns are specific: minimize high-temperature cooking of nitrite-cured meats, monitor sulfite intake if asthmatic, and note that BHA/BHT exposure is declining as manufacturers voluntarily reformulate with tocopherols.