Nutrition · Explainer

Why babies under one can't have honey and how infant botulism happens

Babies under one cannot safely have honey because it can carry bacterial spores that germinate in the infant intestine and produce the nerve toxin responsible for infant botulism. The illness can produce progressive muscle weakness and breathing difficulty (CDC, 2026).

Honey testing, hospital records and surveillance studies address different parts of this risk. Laboratory work identifies spores in some honey samples; patient investigations examine exposure histories, while clinical research describes symptoms, diagnostic uncertainty and recovery. Together, those findings explain the age restriction without attributing every infant botulism case to honey (Harris and Dabritz, 2024).

Glass jar of amber honey and wooden dipper on a kitchen table beside an infant bowl and spoon
A jar of honey beside a small bowl and spoon for infant feeding Editorial illustration

Why babies can't have honey before their first birthday

The first birthday is the boundary used in public health guidance because infant botulism develops mainly in the first year of life. The CDC 2026 prevention guidance excludes honey from the diet before that age. Honey is a recognised source of spores, even though an individual jar is not necessarily contaminated (Centers for Disease Control and Prevention, 2026; Rosow and Strober, 2015).

The relevant hazard is a living bacterium in a dormant form. Infant botulism arises after the spores reach the bowel and conditions permit them to establish themselves there. The process differs from the foodborne illness caused by swallowing toxin that has already formed in another food (Harris and Dabritz, 2024).

Honey as a source of bacterial spores

Honey can acquire C. botulinum spores before reaching a household. Harris and Dabritz (2024) assembled findings from honey surveys and reports that genetically or microbiologically linked environmental material to affected babies. They found spores in a minority of the honey samples examined across those surveys.

In older California investigations, researchers recovered C. botulinum organisms from honey that had been fed to infants with botulism. No preformed toxin was detected by Arnon SS, Midura TF and colleagues (1979). The finding is consistent with a disease in which toxin is generated after intestinal colonisation rather than delivered as an existing ingredient of honey.

Honey's ability to contain spores does not make honey consumption a necessary condition for the illness. Soil and household dust are also investigated as sources. Their presence outside food explains why clinicians consider infant botulism even when no honey appears in the feeding history (Centers for Disease Control and Prevention, 2026; Harris and Dabritz, 2024).

Why the age limit is twelve months

Susceptibility changes as the infant digestive system develops, although a single biological switch at twelve months has not been demonstrated. Harris and Dabritz (2024) describe intestinal colonisation in infants, with particular susceptibility around three to five months and reported illness at younger ages. The developing gut microbial community is one proposed component of that susceptibility.

Healthy adults generally resist colonisation despite encountering spores. In infants, the intestinal ecosystem has not yet established the same resistance, and toxin-producing bacteria can persist long enough to cause illness. These observations support an age-based precaution without defining an exact day when every infant becomes resistant (Harris and Dabritz, 2024).

The restriction concerns every form of feeding in which honey is consumed. The Centers for Disease Control and Prevention (2026) records infant botulism associated with honey-containing pacifiers in Texas. The concern is exposure to spores that can colonise the bowel (CDC, 2026).

Infant botulism symptoms and how the illness appears

Early symptoms can be subtle. The CDC lists constipation, difficulty feeding and a change in crying among possible signs. Progressive weakness provides a more specific neurological pattern, although not every child exhibits every feature (CDC, 2026).

Goldberg and colleagues (2023) reviewed laboratory-confirmed cases from Israel and reported that reduced muscle tone, poor feeding and a weak cry were the common clinical findings. The patients varied in presentation. Their clinical histories show why a symptom list alone cannot determine the diagnosis.

Early changes in feeding and bowel movements

Constipation alone is not diagnostic. The CDC explicitly warns against treating this symptom alone as diagnostic, including when a possible contaminated product is part of the history. More informative findings arise when bowel changes coincide with declining muscle strength or feeding ability (CDC, 2026).

An infant may suck less effectively or lose the strength needed for ordinary feeding. A cry can become faint or altered. These signs reflect interference with the nerves that control muscle movement, but similar signs also occur with other illnesses, including infections and inherited neurological disorders (CDC, 2026).

Illness severity varies between patients. The Israeli series recorded progressive weakness in some babies alongside less typical clinical features in others. A population description of frequent symptoms cannot be turned into a fixed sequence that every affected infant follows (Goldberg et al., 2023).

Muscle weakness and breathing difficulties

Botulinum neurotoxin disrupts acetylcholine release at nerve endings. As a result, muscles cannot receive normal activation signals. The CDC describes weakness that starts in muscles supplied by cranial nerves and can progress to a symmetrical descending flaccid paralysis (CDC, 2026; Harris and Dabritz, 2024).

Visible changes may include drooping eyelids, reduced facial expression and diminished sucking or gag reflexes. Breathing muscles can also become affected. The CDC lists respiratory difficulty and respiratory arrest among possible complications, which distinguishes progressive botulism from an isolated change in feeding or bowel frequency (CDC, 2026).

In Goldberg and colleagues' 2023 Israeli series, classical descending paralysis appeared in 6/8 infants. Two patients needed intubation. Those figures describe the hospital cases included in that small national cohort and cannot predict how severe disease will be in another infant (Goldberg et al., 2023).

How infant botulism develops after spores enter the gut

Clostridium botulinum is a spore-forming bacterium. A spore is a dormant survival form capable of remaining in the environment until conditions permit growth. Infant botulism results when swallowed spores colonise the large intestine and produce neurotoxin locally, as described by the CDC in 2026.

The biological distinction from ordinary foodborne botulism determines how honey is implicated. Foodborne cases involve preformed toxin. Infant botulism is an intestinal infection involving bacterial establishment and toxin production after ingestion (Harris and Dabritz, 2024).

Spore germination and intestinal colonisation

Intestinal colonisation is age-dependent. Harris and Dabritz (2024) discuss immature gut microbiota as a possible explanation for increased susceptibility early in life. The proposed role of intestinal community composition does not establish a particular microbial profile as the cause of every case.

Age patterns also vary. Cases were documented at exceptionally young ages in the 2024 review, while an Israeli hospital cohort had a median diagnostic age of 6.5 months. Those observations do not identify a universal window of susceptibility, especially when infection onset and clinical diagnosis occur at different times (Harris and Dabritz, 2024; Goldberg et al., 2023).

Colonisation can occur after exposure to a source other than honey. Harris and Dabritz describe soil and indoor dust associations along with food sources. Because most cases have no confirmed source, assigning an illness to the last food eaten would exceed the available microbiological evidence (Harris and Dabritz, 2024).

Botulinum toxin and nerve signalling

The toxin blocks communication at cholinergic nerve terminals, where nerves normally release acetylcholine to activate muscle fibres. Harris and Dabritz (2024) describe cleavage of proteins needed for this release. Normal muscle activation then fails.

The range of physical signs reflects the same mechanism in different muscles. Changes in eye movement, facial expression, sucking, swallowing and breathing can emerge as the neurological effects spread. CDC clinical guidance describes a descending pattern of weakness, although early signs may be subtle (CDC, 2026).

Recovery takes time. Neutralising toxin prevents further activity of circulating toxin but does not instantly restore impaired nerve-to-muscle communication. Hospital support may therefore continue while affected functions return (CDC, 2026; Arnon SS, Schechter R et al., 2006).

Where botulism spores in honey and the environment come from

Environmental sampling finds the organism in more places than honey. Soil and sediment can contain viable spores, while dust can bring environmental material into homes. The CDC notes that cleaning does not guarantee removal of spores from household surfaces (Centers for Disease Control and Prevention, 2026).

For infant botulism, exposure evidence has several levels: a product can test positive, a child can have consumed it, and strains recovered from both patient and source can be compared. These observations offer different degrees of evidence about where an infection began. Most individual cases lack a confirmed source (Harris and Dabritz, 2024).

Honey contamination and testing

Harris and Dabritz (2024) surveyed the published testing literature and identified C. botulinum in 173/4,310 (4%) honey samples. The surveys sampled different countries. It measures detection across collected samples, not the probability that a given serving will cause infant botulism.

Some national results differed widely. United States surveys identified C. botulinum in 42/587 (7%) tested honey samples, while three German surveys detected it in 1/354. A negative survey result is bounded by the products and tests examined, rather than a guarantee that every subsequent product will be free of spores (Harris and Dabritz, 2024).

Detection also differs from proof of illness. Arnon SS, Midura TF and colleagues (1979) found organisms in honey supplied to infants with botulism, but no preformed toxin. In Israel, investigators recovered the organism from a honey cake associated with one patient; the reported cake had raw honey added after baking (Goldberg et al., 2023).

Soil, dust and other exposure sources

Harris and Dabritz (2024) compiled 73 environmental surveys from 37 countries conducted between 1919 and 2023. Viable spores appeared in 4,655/36,793 (13%) samples. That denominator covers a range of environmental specimens and cannot be interpreted as the percentage of infants exposed or infected.

Indoor findings were less consistent. A Canadian study in 2023 detected no C. botulinum in 963 household vacuum-cleaner dust samples drawn from 13 cities, according to the 2024 review. Other investigations had previously linked environmental isolates with affected infants, leaving outdoor sources and household transfer relevant despite variable sampling results (Harris and Dabritz, 2024).

A separate food-safety problem, fungal contamination of dried raisins, likewise distinguishes detection in a sampled product from measured harm in a person. The organisms and toxins differ, so those findings do not estimate infant botulism risk.

How infant botulism is diagnosed and treated in hospital

Clinical suspicion begins with progressive weakness rather than with laboratory confirmation alone. The CDC 2026 clinical overview states that symptoms guide the initial assessment and that specialised testing may take days. Negative results do not necessarily exclude the condition (CDC, 2026).

Hospital treatment addresses both circulating toxin and the effects of muscle weakness. Human botulism immune globulin, commonly called BabyBIG, has been evaluated in a controlled trial. The intervention is distinct from general supportive measures such as managing feeding or respiratory problems (Arnon SS, Schechter R et al., 2006; Rosow and Strober, 2015).

Clinical signs and laboratory confirmation

A child with progressive flaccid weakness can resemble a patient with sepsis, spinal muscular atrophy or another neurological illness. The CDC identifies these conditions among important clinical alternatives. Bedside assessment therefore considers the pattern of weakness alongside investigations for other possible causes (Centers for Disease Control and Prevention, 2026).

Definitive infant botulism testing uses a stool or enema specimen and specialised laboratory methods (CDC, 2026). Results are not always available when care begins. Goldberg and colleagues (2023) reported a median interval of 9.5 days from symptom onset to suspected diagnosis in their Israeli cohort, with a range of 4–35 days.

In that cohort, laboratory testing identified toxin-producing clostridia in infants who had varied clinical histories. The study's patient selection depended on confirmed cases, so it could not measure how many children with similar early symptoms finally had another diagnosis (Goldberg et al., 2023).

Hospital care and recovery evidence

The comparison was randomised. Arnon SS, Schechter R and colleagues (2006) compared 59 treated infants with a placebo group, using length of hospital stay as the principal outcome. The mean stay was 2.6 weeks with immune globulin and 5.7 weeks in the control group, a statistically significant difference.

A 2019 Cochrane review judged evidence for earlier recovery with this treatment to be of low or moderate certainty, based on one randomised trial. Only one controlled trial was included. The trial establishes a treatment effect under its study conditions, rather than a guarantee of a particular recovery time for every patient (Chalk et al., 2019).

Rosow and Strober's 2015 clinical review describes the introduction of intravenous immune globulin in 2003 and associated reductions in illness burden and hospital costs. In Israel, Goldberg and colleagues (2023) reported recovery among the patients followed, but the national series was small and used different treatment approaches. Such case outcomes cannot substitute for randomised comparisons.

What population studies reveal about honey exposure and infant botulism

Panditrao and colleagues (2020) examined the descriptive epidemiology of infant botulism in California over its first 40 years after recognition in 1976. Epidemiological estimates depend on where cases are recognised, which laboratory tests are available and whether exposure histories are recorded. An infant may develop botulism without an identifiable honey exposure, and a positive honey sample does not by itself establish a transmission event. These limits become visible when national series are compared (Harris and Dabritz, 2024; Goldberg et al., 2023).

Evidence from epidemiological studies

Goldberg and colleagues' 2023 retrospective multicentre study identified 8 diagnosed cases in Israel during 2007–2021. The reported incidence was 0.3 cases/100,000 live births for that period. Investigators reviewed laboratory-confirmed patients rather than estimating the number of unrecognised infections in the wider population (Goldberg et al., 2023).

Only one infant in that Israeli group had a reported honey exposure. The consumed honey cake also yielded positive laboratory findings, while other cases were investigated for environmental exposures. The contrast shows why a national case count and a honey-attributable case count answer different questions (Goldberg et al., 2023).

Harris and Dabritz (2024) report that approximately 6% of Canadian infant botulism cases from 1979 to 2019 were linked to laboratory-confirmed honey exposure. This figure describes confirmed honey-associated cases, not the total frequency of all possible exposure sources. It does not imply that honey caused the remaining cases or that environmental sources were absent.

Comparison of measured case series

Evidence

Population or samples

Recorded finding

Interpretation

Harris and Dabritz, 2024

4,310 honey samples in published international surveys

173/4,310 (4%) positive

Spore detection, not infant disease incidence

Goldberg et al., 2023

Israel, 2007–2021

8 confirmed cases; 0.3/100,000 live births

National reported incidence in the study period

Arnon SS, Schechter R et al., 2006

59 treated infants and a placebo group

Mean stay 2.6 versus 5.7 weeks

Controlled treatment outcome, not exposure frequency

Dabritz HA et al., 2025

US cases compared with cases elsewhere

Honey exposure 3.8% versus 19.7%

Exposure reporting differs by region

Dabritz HA and colleagues (2025) analysed cases from 2007–2021 and found that honey consumption had been recorded in 19.7% of patients outside the United States compared with 3.8% of US patients. The same analysis reported different ventilation and hospital-stay figures across regions. Surveillance practices and treatment access can affect the comparisons, so these proportions are not geographical estimates of the risk per serving of honey.

A study of DDT in breast milk illustrates a separate measurement question: detecting a substance in a biological sample is not equivalent to determining how an infant acquired a specific infection. The DDT findings do not describe botulism or implicate breast milk as its usual source.

What is proven about honey and infant botulism and what is not

Honey is a documented carrier of C. botulinum spores and a recognised source of infant exposure. Public health guidance excludes it before the first birthday, when intestinal colonisation is a recognised cause of botulism. The precise source of most individual cases remains unknown (Centers for Disease Control and Prevention, 2026; Harris and Dabritz, 2024).

Honey creates an avoidable route of exposure to botulism spores in infancy, although most individual cases cannot be traced to a confirmed source.

The mechanism is established: spores can reach the intestine, toxin-producing bacteria can establish themselves, and the toxin interrupts acetylcholine release at nerves controlling muscles (CDC, 2026). Symptoms range from changes in feeding to respiratory impairment. Laboratory investigations can confirm the disease, but early clinical assessment cannot depend on receiving a positive result (CDC, 2026; Harris and Dabritz, 2024).

What food surveys cannot establish is the chance of illness from a particular jar or individual taste. The worldwide honey survey percentage describes tested samples, while the Israeli and international case figures describe selected populations of patients. None of those numbers is an experimental estimate of risk for a baby given honey (Harris and Dabritz, 2024; Goldberg et al., 2023).

The degree of intestinal susceptibility also varies between infants. Microbial development is a plausible part of the age pattern, but exposure, colonisation and clinical illness remain separate events. An age boundary in guidance is therefore a public health precaution grounded in observed disease patterns, not proof that every infant is equally susceptible until a precise date (Rosow and Strober, 2015; Harris and Dabritz, 2024).

Source attribution remains the main evidential gap. More investigations matching a patient's organisms with environmental and food isolates would clarify which routes account for currently unexplained cases. Until then, the association with honey supports an age restriction, while the absence of honey consumption cannot exclude infant botulism (Harris and Dabritz, 2024; Centers for Disease Control and Prevention, 2026).

Frequently asked questions

Can a baby get botulism from cooked or baked honey?

A food containing honey may still pose an infant botulism exposure concern because bacterial spores can survive heating that does not reliably destroy them. In one confirmed Israeli case, raw honey had been added to a cake after baking. That case documents exposure but cannot establish the safety of other preparation methods (Goldberg et al., 2023).

Is pasteurised honey safe for babies?

No distinction between pasteurised and unpasteurised honey appears in the CDC's under-one age restriction. Infant botulism depends on viable spores reaching the bowel, and the presence of spores cannot be established from a product label alone. Detection studies also show that contamination is variable among sampled products (Centers for Disease Control and Prevention, 2026; Harris and Dabritz, 2024).

Can a single taste of honey cause infant botulism?

A probability for one taste is not established by surveillance and honey-testing studies. Spore exposure, intestinal colonisation and illness are separate events. The CDC's guidance excludes honey throughout the first year, but does not state that every exposure produces infection or that a small quantity is safe (Centers for Disease Control and Prevention, 2026).

Can infant botulism occur without eating honey?

Yes. Honey is a documented source of spores, but most infant botulism cases are not assigned to a laboratory-confirmed source. Soil and household dust have also been investigated. In the Israeli series from 2007–2021, only one of eight infants had a recorded honey exposure (Goldberg et al., 2023; Harris and Dabritz, 2024).

Does honey eaten by a breastfeeding mother put her baby at risk?

Infection follows spores entering the infant intestine (CDC, 2026). Available reports concern infant exposure, not maternal consumption. The reviewed studies do not establish breast milk as a route for transmitting spores from an adult's honey consumption to an infant (CDC, 2026; Harris and Dabritz, 2024).

Sources

  1. Harris RA, Dabritz HA. Infant Botulism: In Search of Clostridium botulinum Spores. Current Microbiology. 2024;81:306. DOI 10.1007/s00284-024-03828-0.
  2. Goldberg B, Danino D, Levinsky Y, et al. Infant Botulism, Israel, 2007–2021. Emerging Infectious Diseases. 2023;29(2):235–241. DOI 10.3201/eid2902.220991.
  3. CDC. Clinical Overview of Infant Botulism. 29 September 2026. https://www.cdc.gov/botulism/hcp/clinical-overview/infant-botulism.html
  4. Centers for Disease Control and Prevention. Botulism Prevention. 26 February 2026. https://www.cdc.gov/botulism/prevention/index.html
  5. Rosow LK, Strober JB. Infant botulism: review and clinical update. Pediatric Neurology. 2015;52(5):487–492. DOI 10.1016/j.pediatrneurol.2015.01.006.
  6. Arnon SS, Schechter R, Maslanka SE, Jewell NP, Hatheway CL. Human Botulism Immune Globulin for the Treatment of Infant Botulism. New England Journal of Medicine. 2006;354:462–471. DOI 10.1056/NEJMoa051926.
  7. Chalk CH, Benstead TJ, Pound JD, Keezer MR. Medical treatment for botulism. Cochrane Database of Systematic Reviews. 2019;(4):CD008123. DOI 10.1002/14651858.CD008123.pub4 
  8. Dabritz HA, Chung CH, Read JS, Khouri JM. Global Occurrence of Infant Botulism: 2007–2021. Pediatrics. 2025;155(4):e2024068791. DOI 10.1542/peds.2024-068791.
  9. Arnon SS, Midura TF, Damus K, Thompson B, Wood RM, Chin J. Honey and other environmental risk factors for infant botulism. Journal of Pediatrics. 1979;94(2):331–336. DOI 10.1016/S0022-3476(79)80863-X.
  10. Panditrao MV, Dabritz HA, Kazerouni NN, et al. Descriptive Epidemiology of Infant Botulism in California: The First 40 Years. Journal of Pediatrics. 2020;227:247–257.e3. DOI 10.1016/j.jpeds.2020.08.013.
Elias R. Westbrook

Elias R. Westbrook — author

Elias R. Westbrook is academic background in human nutrition and nutritional sciences, with a focus on food composition, nutrient intake, dietary patterns, and the relationship between foods and nutritional health. For Cabdirect, this persona is intended to co...

This article is for general information and is not medical advice. It was reviewed for accuracy by a qualified clinician; decisions about your health should be made with your own doctor.

Related

About this section