Clinical manifestations

Presentation and course

Infant botulism usually occurs in previously healthy, full-term infants, most often between 6 weeks and 6 months of age (range of reports 6 days to 1 year). Symptoms typically evolve over a period of days to even a week, with a clinical nadir reached at 1 to 2 weeks. Constipation and cranial nerve palsies (commonly manifesting as drooling, poor feeding, or a weak cry) are often the earliest symptoms, followed by a flaccid descending paralysis (73). A case series revealed poor feeding and lethargy to be the most common symptoms prompting presentation (23). In a chart review of 44 infants with infant botulism, 43 of the infants were admitted for poor feeding, generalized weakness, and hypotonia with a range of symptoms from 0 to 22 days prior to admission (72).

Affected infants experience the abrupt or gradual onset of systemic weakness and bulbar dysfunction, manifested by inactivity, hypotonia, weak cry, poor suck, and reduced oral intake. Infants with infant botulism appear “lethargic” or apathetic to their parents or care providers because of their ptosis, inactivity, and paucity of facial expression. Most infants have constipation, and several days may pass between bowel movements. Systemic muscle weakness and respiratory distress become major features as the disease progresses. Though diaphragmatic function is thought to be preserved until 90% involvement (23), respiratory failure can occur abruptly. A case report of an infant diagnosed with respiratory syncytial virus but persistent respiratory failure, cranial nerve palsies, and subsequent stool + for botulinum toxin A emphasizes the importance of considering infant botulism within the differential diagnosis for these signs and symptoms despite other confirmed diagnoses (59).

Botulinum toxin does not cross the blood-brain barrier, and there are no associated central nervous system symptoms, nor consequences.

Examination of the infant with botulism reveals ptosis, facial weakness, feeble cry, poor suck, reduced or absent gag reflex, poor head control, diffuse hypotonia, weakness, and reduced or absent muscle stretch reflexes (70). The pupils are usually dilated and react sluggishly to light (22). Sluggish pupil response, however, may take continually repeating the exam for 2 to 3 minutes to elicit constrictor muscle fatigability. Fever is characteristically absent.

Antitoxin effect from BabyBIG^® is reported to neutralize botulinum toxin for 6 or more months (08).

Prognosis and complications

Many infants with botulism require mechanical ventilation due to respiratory failure or difficulty handling oral secretions. Prior to the availability of botulism immune globulin, the duration of assisted ventilation was 3 or more weeks (62). In a randomized controlled trial of botulism, immune globulin treatment reduced the mean duration of mechanical ventilation by 2.6 weeks (08). Clinical guidelines published by the CDC in 2021 recommend administration of botulinum antitoxin to patients with suspected botulism as early as possible in the course of illness. The greatest benefit accrues to those who receive it within the first 2 days of illness onset (60). Infants require nasojejunal feedings until they recover bulbar function and can appropriately feed and protect their airways (70).

Relapse occurs in approximately 5% of infants with botulism, usually within 2 weeks of recovery from the initial illness (30).

Clinical recovery occurs over a period of weeks to months, but full clinical recovery is expected.

Clinical vignette

A previously healthy 4-month-old male infant was brought to the emergency department for lethargy and reduced oral intake. There was no family history of serious systemic or neurologic disease. Presentation in the emergency department revealed an inactive infant with normal vital signs, afebrile. General examination showed no heart murmurs, organ enlargement, rash, or dysmorphic features. Neurologic examination showed bilateral ptosis, weak cry, diminished gag, and diffuse hypotonia with poor head control. Pupils were midrange and sluggishly reactive to light. Muscle stretch reflexes were present but diminished; sensation was intact to light touch and noxious stimuli. His parents remarked that he normally had excellent head control and previously cried vigorously when examined. His last bowel movement was 5 days earlier.

Laboratory studies included normal complete blood cell count, urinalysis, and serum electrolytes, calcium, and magnesium. Electrodiagnostic studies showed a 40% increment during rapid stimulation. Stool studies were subsequently positive for both C botulinum organism and toxin (type A). He was admitted to the hospital, treated with botulism immune globulin (BabyBIG^®), and provided with aggressive supportive measures; he recovered well.

Biological basis

Etiology and pathogenesis

Infant botulism results from toxin produced by swallowed spores most commonly of Clostridium botulinum and rarely of C butyricum or C baratti. Etiology of infant botulism is distinctly different from food-borne botulism (the ingestion of a preformed toxin) or wound botulism (from inoculation of the bacteria into a wound). Ingestion of honey is a common association, but spores are also found in aerosolized particles from soils (such as construction sites or rural locations) and house dust.

Botulinum toxin is formed under anaerobic conditions by Clostridium species, ubiquitous spore-forming organisms found in soils and households. Seven serotypes (types A through G) of toxin have been identified (51; 67). Types A, B, E, and F are associated with human disease (38); types A and B produce the majority of cases of infant botulism (50; 51). Luquez and colleagues reported a case of infant botulism due to Clostridium botulinum type E (49). From 1976 to 2016, C baratii type F caused more than 0.5% of cases occurring mostly in very young infants with a more rapid and severe illness compared to types A and B (36). In 2014, Barash and Arnon reported the first new strand of Botulinum neurotoxin (type H) in over 40 years, isolated from a fecal sample of an infant with botulism (10).

Rarely, enteric colonization during broad spectrum antibiotic therapy in adults leads to clinical botulism (12; 51; 33; 67).

Once ingested, C botulinum spores can colonize the large intestine of infants, multiply, and produce toxin (53; 67). Because of competitive inhibition by other microorganisms, C botulinum is not part of the normal intestinal microflora (03). Clostridia sp are excluded from the gut of breastfed infants, but the addition of solid foods can result in colonization (29). This accounts for cases of infant botulism in breastfed infants who become colonized with C botulinum during introduction of other foods (03; 47). Once absorbed, the toxin is carried hematogenously to the neuromuscular junctions (51). Botulinum toxin in infant botulism has not been reported to cross the blood-brain barrier into the CNS.

The C botulinum toxin, a complex protein similar to the tetanus toxin, is among the most potent of neurotoxins (54). It is synthesized as a large 150 kd protein and consists of two disulfide-bonded fragments. The larger 100 kd heavy chain fragment of the botulinum toxin binds irreversibly to specific receptors in the axon terminus of cholinergic motor neurons, leading to an inability of motor neurons to release acetylcholine from the presynaptic terminal in response to an action potential (66). Structural analysis reveals binding of botulinum toxin type B to synaptotagmin II and specific ganglioside receptors (20). Data have indicated that antigenic peptides at the heavy-chain terminal contain the binding sites to the receptor synaptotagmin II, allowing for botulinum toxin type B binding (65). After binding, the toxin is then translocated and exerts intracellular effects to prevent the release of acetylcholine into the neuromuscular junction.

Various host factors also seem to be important as C botulism spores are more commonly ingested than clinical disease produced (15). A report of a 4-month-old infant with H1N1 influenza and subsequent diagnosis of infant botulism may support the theory of impaired host response; alternatively, this case could have already had exposure to the botulism spores prior to development of influenza as the botulism incubation period in the gut ranges from 3 to 30 days (40). The effects of toxin binding at the neuromuscular junction and other cholinergic nerve terminals, such as the intestine, produce the clinical manifestations of muscle weakness and constipation (72).

The precise biological events that lead to clinical recovery in human infants are unclear because botulinum toxin persists in stool samples during and after resolution of clinical symptoms (06; 47). In fact, a case report presented a 3-month-old infant with persistent excretion of C botulinum cells in stool samples for 7 months following diagnosis of type A infant botulism despite symptoms resolution occurring within 2 months (25). Recovery appears to require axonal transport of proteins from the perikaryon and regeneration of new motor end plates (51). Morphologically, recovery is associated with new sprouting from the presynaptic terminal in the neuromuscular junction. Given the need for regeneration of terminal motor neurons and formation of motor end plates, complete clinical recovery is expected to take weeks to months.

Epidemiology

In 2018, a total of 242 cases of infant botulism were reported, which accounted for 67% of all total cases of botulism reported to the National Botulism Surveillance Summary System (19). Of those, 43% were type A toxin, 54% type B toxin, 1% Ba toxin, 2% Bf toxin, and <1% F toxin. Affected infants ranged in age from 0 to 10 months (median of 4 months). Infant botulism has been recognized in all regions of the United States, 31 states (18; 67; 19), and on all continents except Africa (04; 44). Cases of infant botulism in the United States cluster in certain geographic regions, with California (23%) and Pennsylvania (17%) reporting the most (19). Although cases can occur year-round, the prevalence tends to be highest in the summer and fall (05; 70). Type A is found predominately west of the Mississippi River and type B predominately east of the Mississippi River. Type F has been reported in various regions of the United States (36).

Globally, infant botulism is thought to be under-recognized and under-reported. Evaluating the global occurrence from 1976 to 2006, 26 countries on five continents, in addition to the United States, had reports of infant botulism. The most frequent cases of infant botulism outside of the United States occur in Argentina, Australia, Canada, Italy, and Japan (44). In the United Arab Emirates, there is a case report of an infant with “double toxin” botulism type Ba (26). In Italy, four of the 23 cases reported between 1984 and 2005 were secondary to C Butyricum, type E neurotoxin and 73% after reported honey ingestion (27). France has noted an increase in reports of infant botulism over the past 6 years, citing seven cases between 2004 and 2009 (43). Globally, type A toxin was reported most commonly, in nearly 85% of cases, followed by type B toxin (10%), and rarely type E toxin (44).

Although cases can cluster geographically (70), point source outbreaks of infantile botulism do not occur. There are no reported cases of infant-to-infant transmission of infant botulism. Exposure to the organism is the major risk factor, as indicated by an increased relative risk among early cases following dietary exposure to honey (odds ratio 9:8) or corn syrup (odds ratio 5:2), items that can contain C botulinum spores (68). Exposure to these foods explained 16% to 30% of the early cases (07; 68), but honey and corn syrup have been implicated in less than 10% of the recently described cases in the United States (70; 08). C botulinum spores have been detected in 7.5% of Chamomile tea samples in Argentina, often used in herbal remedies, which could be a potential vector for infant botulism (14). PCR analysis of a random collection of honey samples in Turkey revealed 2.6% positive samples, three type A and one type B toxin, suggesting continued potential risk for infant botulism exposure (35). In Ireland, infant botulism (C butyricum neurotoxin type E) was associated with exposure from pet terrapins (64).

Risk factors may vary according to the age of the infant. Rural residence and exposure to aerosolized soils are risk factors (71; 70; 05; 68). Longstanding constipation and breastfeeding are controversial risk factors. As constipation is an early symptom, it is difficult to attribute a causal role. In infants exclusively breast fed, the changes in intestinal microflora that occur when formula or solid foods are introduced may play a role in the ability of C botulinum to colonize the infant's colon (70). The majority of the Pennsylvania cases reported by Long and colleagues occurred within 4 weeks of the initial introduction of non-breast milk foods (47). Reported findings of soil-dwelling Clostridia sporogenes and Clostridia butyricum in samples of powdered infant formula indicate the risk of the potential presence of Clostridia botulinum in powdered infant formula products (11). Neither botulism bacteria nor toxin are able to pass through breast milk, thus, maternal ingestion of honey or maternal infection with food-borne botulism are not thought to be risk factors in the development of infant botulism.

Infant botulism was once postulated to be a factor in the pathogenesis of sudden infant death syndrome (SIDS) (51). In one large series, toxin or organism was recovered in 4.9% of 211 cases of sudden infant death (05). However, both disorders occur in infants between the ages of 2 and 6 months, suggesting that the relationship may be coincidental rather than causal. An Australian study failed to identify C botulinum in cases of SIDS (16).

Prevention

Given the strong, historical association with infant botulism, honey and corn syrup should be avoided in young infants. Primary care physicians must be aware of the clinical features of infant botulism because early and appropriate diagnosis is critical to the effective management of the disorder and avoidance of potentially fatal respiratory complications.

Benjamins and colleagues surveyed parents presenting to a local county hospital pediatric clinic on use of and perceived risk associated with honey pacifier use in infants based on clinical staff observations of frequent use of these pacifiers (13). Nearly 400 parents responded to the survey, and11% of this population reported providing honey pacifiers to their infants, greatest in the Hispanic population within this sample; 80% were unaware of any risks from exposing their infants to honey. This study emphasizes the importance of continued focus on prevention through parent education on risks associated with honey, especially in infants.

As excretion of botulinum toxin or cells may remain positive months after clinical resolution, good hand hygiene after each diaper change and proper, rapid disposal of diapers from infants with botulism is imperative to prevent contamination of others.

Differential diagnosis

Infants with botulism have clinical features that suggest bacterial sepsis or meningitis. Other disorders that can be associated with acute weakness include electrolyte disorders, such as extreme hyponatremia, hypokalemia or hyperkalemia, hypomagnesemia or hypermagnesemia, hypothyroidism, hypocalcemia or hypercalcemia, thiamine deficiency, and inborn errors of metabolism, especially disorders of the urea cycle, organic acid metabolism, or fatty acid oxidation defects.

Other considerations include spinal muscular atrophy, infantile Guillain-Barré syndrome, tick paralysis, poliomyelitis, mitochondrial disorders, glutaric aciduria type II, and infantile Pompe syndrome. Spinal muscular atrophy is suggested by areflexia, tongue fasciculations, and chronicity of signs and symptoms. Furthermore, spinal muscular atrophy typically spares the ocular muscles. Both spinal muscular atrophy and Guillain-Barré syndrome can be distinguished from infant botulism by distinctive electrophysiologic features. Although the Miller Fisher variant of Guillain-Barré syndrome involves cranial nerve palsies, this disorder can be distinguished by CSF analysis, nerve conduction studies, and EMG findings. Eradicated from the United States and most regions of the world, poliomyelitis can be distinguished from infant botulism by antecedent gastroenteritis, fever, signs of aseptic meningitis, and the presence of asymmetric weakness. Infants with mitochondrial disorders have elevated serum lactate and often distinctive abnormalities on brain MRI. It is also important to consider toxic exposures such as organophosphate or heavy metal poisoning.

Rarely, unrecognized congenital myasthenia gravis mimics infant botulism. Like infants with botulism, infants with myasthenia gravis have ocular involvement (rarely with pupillary abnormalities, however) and bulbar weakness. The response to edrophonium chloride or neostigmine can be confusing, as infants with botulism may transiently improve after administration of these agents. The response to rapid repetitive stimulation of motor nerves (24) and the recovery of C botulinum toxin from the stool samples can distinguish these two disorders.

Five categories of clinical mimics were described based on 681 cases from 1992-1997. These “mimics” occurred in 4.7% of infants in whom the laboratory diagnosis of infant botulism was not confirmed and include: (1) spinal muscular atrophy, (2) metabolic disorders, (3) other infectious diseases (ie, enterovirus encephalitis, RSV bronchiolitis, viremia), (4) miscellaneous (ie, Miller Fisher variant of Guillain-Barré syndrome, presumed Lambert Eaton syndrome, cerebral atrophy and infarctions, idiopathic spinal epidural hematoma, diaphragmatic paralysis and central demyelinating disease), and (5) probable infant botulism (28). Reported potential surgical mimics include Hirschsprung disease (55) and presentation of an acute abdomen (58).

In a follow-up report by Khouri and colleagues in 2018, 1226 patients were treated with BIG-IV from 2005 to 2015, of which 76 (6.2%) clinical mimic illnesses were identified (42). These were also divided into the same categories listed above. The distribution revealed (1) probable infant botulism lacking confirmatory testing (26.3%); (2) spinal muscular atrophy (19.7%); (3) miscellaneous (15.8%) (ie, abnormalities in gray/white matter and corpus callosum, acute disseminated encephalomyelitis, transverse myelitis, Chiari malformation, cystic fibrosis, hypovitaminosis A, hemophilia A, cervical epidural hemorrhage, hypothyroidism, nemaline rod myopathy, and Miller-Fischer variant or pharyngeal-cervical-brachial variant of Guillain Barre Syndrome); (4) metabolic disorders (11.8%); and (5) other infectious diseases (10.6%) (human metapneumovirus, parainfluenza, parechovirus, roseola, polio-like enterovirus, likely resolved sepsis). No alternate diagnosis was established in 15.8%, and these were, therefore, categorized as undetermined.

Diagnostic workup

Infants with botulism have C botulinum spores and toxin in their stools. Detection of toxin confirms the diagnosis of infant botulism, whereas detection of organisms only establishes colonization. Affected infants commonly have reduced stooling, necessitating saline enemas for effective stool sampling. Stool can remain positive for culture and toxin for extended periods, although infants may clear the organisms within 1 month of diagnosis (39). It is recommended to collect stool specimens for laboratory confirmation of the clinical diagnosis of botulism as soon as possible because toxin levels decrease over time (60).

The current standard of organism and toxin detection is through a toxin neutralization mouse bioassay. Polymerase chain reaction methods can be used to detect toxin, but the sensitivity and specificity of polymerase chain reaction have not yet been established (69; 41; 01). Glyco-qPCR detection is another rapid analysis technique being explored (45). Several other diagnostic techniques are being explored, including ELISA, endopeptidase activity assay, biosensors, cell-based assay, mass spectrometry, lateral flow, rolling circle amplification, and other techniques reviewed (46; 17; 37).

Potentially life-threatening systemic or CNS infections should be considered in infants with acute weakness and hypotonia. Appropriate cultures of blood, urine, and CSF should be obtained. Serum electrolytes, calcium, magnesium, thyroid, lactate, and ammonia should also be considered. Specimens for toxicology and metabolic studies may be appropriate, depending on the clinical history. A spinal tap is often indicated, but this procedure must be performed carefully because prolonged neck and trunk flexion may provoke respiratory failure.

Imaging of the brain and spine is often not indicated, but when reported, is typically normal/negative. However, Good and colleagues reported a case of a 5-month-old infant that presented clinically with features of infant botulism (ultimately confirmed with mouse bioassay), but had abnormalities on brain MRI (areas of white matter diffusion restriction) and spine MRI (symmetrical cervical nerve root enhancement) (31). This infant responded rapidly to BabyBIG^®, and the authors noted that neuroimaging findings should not preclude treatment in patients with a clinical presentation consistent with infant botulism.

Although largely a clinical diagnosis, infant botulism can be supported by electrophysiologic studies. An incremental response to rapid (20 or 50 Hz) repetitive stimulation is the most sensitive and specific finding (70). Slower rates of repetitive stimulation yield variable responses and confusing results (24). Needle electromyography is also sensitive but less specific, and many infants have short-duration, low-amplitude motor unit potentials and abnormal spontaneous activity (22; 24). False negative or positive EMG results can be observed (32; 70). A review of 14 patients with infant botulism revealed brief, small amplitude potentials as the most common EMG abnormality (72).

Management

Human botulism immune globulin (BIG-IV, known as BabyBIG^®) shortens the clinical course of infant botulism as measured by the durations of intensive care, mechanical ventilation, hospitalization, and intravenous tube feedings (63; 02; 70; 08; 73; 60).

The Committee on Infectious Diseases of the American Academy of Pediatrics recommends that BabyBIG^® be administered as soon as the diagnosis of infant botulism is suspected, without waiting for laboratory confirmation (02; 73). BabyBIG^® can be obtained from the California Department of Health Services (phone 510-231-7600). Information can also be obtained at www.infantbotulism.org.

The half-life of BabyBIG^® is approximately 28 days, and a single infusion is reported to neutralize toxin for 6 months (08). Although botulism immune globulin is typically given within 2 weeks of symptoms and is most effective within 72 hours (08), there is report of effective, rapid response as late as 23 days after presentation (26). Nevertheless, BabyBIG^® should be given as soon as infant botulism is clinically suspected, ideally within 2 days (60).

The primary adverse event associated with BabyBIG® was a transient erythematous rash. No other adverse events were more common in the BabyBIG® group when compared to the placebo-treated group (08; 73). The Cochrane Collaboration released a report indicating good evidence to support the use of BabyBIG® in the treatment of infant botulism (21).

As the cost of BabyBIG^® limits use in other countries, equine botulinum antitoxin has been documented as a safe, effective alternative (74).

Infants with botulism require meticulous supportive care (02). The major threat to infants with botulism is respiratory failure. Consequently, affected infants should be monitored closely in an intensive care setting; they should be intubated and ventilated when signs of airway compromise, bulbar dysfunction, or respiratory insufficiency arise. Sedation for scans or other procedures should be given very cautiously. Untreated infants with botulism may require ventilation for several weeks (75; 08).

Infants may experience altered autonomic function, including sudden unexplained alterations in heart rate, blood pressure, or skin color. Symptoms of autonomic dysfunction are usually mild or transient and rarely require specific therapy (47). Urinary retention may require intermittent catheterization. A small number of infants experience the syndrome of inappropriate antidiuretic hormone secretion, especially early in the disorder after being placed on positive pressure ventilators. This responds to fluid restriction and usually resolves spontaneously.

Constipation responds to stool softeners, digital manipulation, and enemas. Despite their extreme constipation, most infants with botulism tolerate enteral feedings well. Because of the potential for aspiration, nasojejunal feedings are preferred over nasogastric. Most infants can be adequately nourished by continuous nasojejunal feedings, beginning as early as 3 days after presentation (62). Feeding should not be resumed until the infant has normal gag, suck, swallow, and airway protection.

Antibiotics may result in increased toxin release, and aminoglycosides further impair release of acetylcholine, both worsening neuromuscular blockade (72). If possible, broad spectrum antibiotics should be avoided. Despite transient improvement with intravenous edrophonium or intramuscular neostigmine, acetylcholinesterase inhibitors do not have sustained benefit.

A case report suggests an association between thiamine deficiency and severity of infant botulism symptoms, as thiamine is required for acetylcholine synthesis, and recommends consideration of thiamine supplementation (61).

Finally, after administration of BabyBIG^®, immunizations should be delayed until the infant has regained full muscle strength/tone, and live viruses should be delayed 5 months (www.infantbotulism.org).

Outcomes

Unless respiratory failure and death occur during the acute stage of the disorder, infants with botulism recover completely. During the acute phase, nearly all infants will have compromised respiratory function, and approximately 50% of patients presenting with infant botulism require mechanical ventilation (70; 08). Two factors that have been reported to be associated with respiratory decompensation include use of aminoglycoside antibiotics, which further decreases release of acetylcholine, and neck flexion for procedures (15).

The syndrome of inappropriate antidiuretic hormone can be observed in some infants (47), necessitating periodic assessment of serum electrolytes. Secondary infections, including otitis media and pneumonia, are common in severely ill infants requiring various degrees of life support (62). Urinary retention and constipation are frequent associated complications. Relapse occurs in approximately 5% of infants with proven botulism, usually within 2 weeks after initial recovery (30). However, after administration of BabyBIG^®, neutralizing toxin persists for 6 months or more, thus, making recurrence unlikely. Parents should be reassured that their infants will recover completely, despite the prospects of a prolonged hospitalization.

Special considerations

Anesthesia

In the nonintubated patient with botulism, sedation has a high risk of respiratory suppression. Neuromuscular blocking agents should be avoided if possible. Intubated infants with botulism often require artificial ventilation for extended periods.