Detection and characterization of Clostridium botulinum isolated from powdered infant formula

Abstract

Introduction:

As a part of an ongoing investigation of the 2025 multistate outbreak of infant botulism linked to consumption of powdered infant formula (PIF), various samples – including unopened containers of PIF and base powder (bulk PIF before packaging) – were analyzed for the presence of Clostridium botulinum.

Methods:

Samples were screened using a tiered analytical approach: anaerobic enrichment followed by real-time PCR targeting botulinum neurotoxin-associated genes, confirmatory PCR, amplicon sequencing, colony isolation, and whole-genome sequencing (WGS). C. botulinum was detected in both the finished product and base powder.

Results:

Genomic analysis revealed genetic identity between isolates from one finished product lot and a base powder, while an isolate from another lot was distinct. Importantly, detection occurred in samples with non-detectable sulfite-reducing clostridia (SRC).

Conclusion:

Results of this study demonstrate that indicator-based screening (such as SRC enumeration) even if it had been in place before the outbreak, would not have prevented its occurrence. Positive linkage of an infant botulism outbreak to PIF should compel manufacturers to recognize C. botulinum as a hazard reasonably likely to occur in certain ingredients, necessitating the design and implementation of specific preventative controls.

1 Introduction

Clostridium botulinum is an obligate anaerobic, spore-forming bacterium capable of producing botulinum neurotoxins (BoNT), which are among the most potent biological toxins known. These neurotoxins cause botulism, a rare but severe paralytic illness that presents a particular risk to infants, whose immature gut microbiota and physiological conditions allow intestinal colonization and in situ toxin production (Centers for Disease Control and Prevention (CDC), 1998; Lindström and Korkeala, 2006; Harris and Dabritz, 2024). Infant botulism is most associated with environmental exposure to spores originating from soil and dust; however, additional exposure pathways remain of concern given the extreme toxicity of BoNT and the low infectious dose required for disease (Hauschild and Dodds, 1993; Peck, 2009).

PIF has historically been considered a potential vehicle of concern for spore-forming bacteria because of its low water activity, extended shelf life, and the inherent resistance of clostridial spores to desiccation and other environmental stresses (American Public Health Association, 2015; Hauschild and Dodds, 1993; Doyle and Buchanan, 2013). Despite the theoretical considerations, the prevailing view and the advice given to the industry in recent years has been that C. botulinum is not a hazard reasonably likely to occur in PIF manufacturing.

In a 2004 Joint FAO/WHO Expert Meeting on Microbiological Risk Assessment, C. botulinum was classified as a microorganism capable of causing illness in infants but not identified as a hazard in PIF (Joint FAO/WHO Expert Meetings on Microbiological Risk Assessment (JEMRA) and World Health Organization, 2004). Consequently, C. botulinum is not regarded as a hazard in the Codex Code of Hygienic Practice for Powdered Formulae for Infants and Young Children (Codex Alimentarius Commission and Food and Agriculture Organization of the United Nations, 2008). Building on this assessment, the International Commission on Microbiological Specifications for Foods (ICMSF) (2014) recommended the use of SRC enumeration as a general indicator of anaerobic spore contamination, rather than routine testing for clostridial spores or C. botulinum, reflecting the presumed low prevalence of the organism in infant formula.

Nevertheless, the severe clinical consequences of botulism and the heightened vulnerability of the infant population necessitate continued vigilance. Within this context, the detection, isolation, and characterization of C. botulinum from infant formula–associated matrices provide valuable insights into contamination pathways, organism persistence, and the limitations of current detection strategies (Peck, 2009; Popoff and Bouvet, 2013).

In 2025, a multistate outbreak of infant botulism in the United States prompted recalls of PIF products¹. As part of the associated investigation, IEH Laboratories and Consulting Group was engaged to screen, detect, and characterize C. botulinum from finished products, using a tiered analytical approach combining anaerobic enrichment, molecular screening, culture confirmation, and genomic analysis.

2 Materials and methods

2.1 Enumeration assays

A 25 g portion of sample was enumerated for aerobic and anaerobic mesophilic spores according to the Compendium of Methods for the Microbial Examination of Foods (Fifth edition). Sulfite reducing clostridia were enumerated according to ISO 15213-1:2023 (2023).

2.2 Enrichment and qPCR

A 100 g portion of the sample was enriched for botulinum-producing clostridia (spores and vegetative cells) by germination and growth in tryptone-peptose-glucose-yeast extract (TPGY) broth under anaerobic conditions following ISO/TS 17919:2013 (2013) protocol with minor modifications such as larger sample size. All samples were subjected to heat shock at 70–80 °C for 10–12 min before enrichment. After enrichment, a 1 mL aliquot of enrichment media was subjected to DNA purification using silica coated magnetic beads and the DNA eluted in 50 μL of sterile deionized water. A 2 μL aliquot of eluted DNA was tested using a C. botulinum BoNT/A and BoNT/B Real-Time PCR assay. The assay uses primers and probes for BoNT/A and BoNT/B as specified in ISO/TS 17919:2013 (2013) Annex C. The assay targets BoNT/B (FAM), BoNT/A (HEX), and C. spp. 16S (ROX) ribosomal RNA genes, as well as an internal control target in the Cy5 channel. PCR cycling was as follows: 95 °C, 5 min; 40 cycles of 95 °C, 20 s, 55 °C, 20 s, 72 °C, 20 s.

2.3 Confirmation of the presence of C. botulinum

Presumptive positive enrichment cultures were subjected to multiple PCR amplifications using the FDA botA983 primer set (U.S. Food and Drug Administration (FDA), 2025), an in-house botA782 primer set, and the FDA botB492 primer set (U.S. Food and Drug Administration (FDA), 2025). PCR amplicons were purified using silica-based spin columns and sequenced on a Flongle flow cell using the MinION Mk1C platform (Oxford Nanopore Technologies), following manufacturer’s instructions.

2.4 Isolation and whole genome sequencing of C. botulinum

Confirmed positive enrichment cultures were streaked on Botulinum Selective Media (BSM) and incubated under anaerobic conditions to obtain isolated colonies. Representative colonies were screened by in-house multiplex PCR assays targeting BoNT/A, BoNT/B, fldB and ntnh genes. PCR-positive colonies were subsequently enriched overnight in TPGY broth. Genomic DNA was extracted from overnight cultures using silica-based spin columns, followed by library preparation with Nextera XT DNA Library Preparation Kit and sequencing using the MiSeq v2 Reagent Kit on the MiSeq platform (Illumina, CA, United States).

2.5 Sequencing analysis

The raw sequencing reads were checked using FastQC for Quality Control before de novo assembly using the SPAdes assembler. Species identity was confirmed using the Type (Strain) Genome Server (TYGS). Single nucleotide polymorphism (SNP) analysis was performed using the FDA CFSAN SNP Pipeline to assess genetic relatedness among isolates.

3 Results

Of the three lots of finished product initially tested as part of the outbreak response, 2 lots had presumptive positives (BoNT/A) – 4 of 12 samples from the first lot and 2 of 12 samples from the second lot (Table 1).

Of the 9 lots of base powder initially tested, 3 lots had presumptive positives – 2 of 8 samples from one lot and 1 of 8 samples each from the other two lots (Table 2).

All six presumptive positives from finished product (Table 1) and two presumptive positives from base powder confirmed positive for C. botulinum by amplicon sequencing of the neurotoxin genes.

Confirmed positive enrichment samples were streaked for isolation. Four colonies from lot #1233393-005 and one colony from lot #1233392-004 of finished product were isolated. One colony from lot #1233418-066 of base powder was also isolated.

Bacteria from isolated colonies were subjected to whole genome sequencing and compared to each other as well as four whole genome sequences from NCBI using SNP analysis (Table 3). The four isolates taken from finished product lot no. 1233393-005 are identical to each other but very different (>400 SNPs) as compared to the isolate from finished product lot no. 1233392-004. The two isolates from base powder are identical in sequence to the four isolates from finished product lot #1233393-005. None of the isolates were a genetic match to the NCBI strains listed in the table as well as in a more comprehensive search of all available genome sequences in NCBI.

4 Discussion

The 2025 United States infant botulism outbreak, linked to a brand of infant formula with 51 cases reported at the time of this submission, is a watershed event. It shifts the risk assessment for C. botulinum in PIF from a theoretical hazard to one that is “reasonably likely to occur.” Currently, PIF and specific ingredients, such as dairy materials, are not subjected to pasteurization processes capable of delivering a 6-log reduction of heat-resistant C. botulinum spores.

Furthermore, testing for C. botulinum is extremely challenging, particularly with regards to culture confirmation. Real-time PCR screening of sample enrichments revealed that non-target clostridial species can be present at concentrations six orders of magnitude (10⁶) higher than C. botulinum. This resulted in Cq values more than 20 cycles earlier than the target (data not shown). In this study, examining more than 1,500 colonies was required to find a single positive C. botulinum colony, highlighting the critical importance of molecular approaches for confirmation.

Following the outbreak and subsequent recall, several groups promoted SRC testing as a primary control measure. However, our evidence highlights significant risks in this approach. The recovery of C. botulinum from finished products – confirmed by real-time PCR and genomic analysis – demonstrates the limitations of indicator-based screening. Notably, these findings occurred in samples where SRC were not detected, indicating that SRC enumeration did not reliably correlate with the presence of toxigenic C. botulinum. Similar shortcomings of indicator organisms have been noted previously (Hauschild and Dodds, 1993; International Commission on Microbiological Specifications for Foods (ICMSF), 2014).

This observation is consistent with the biological heterogeneity of the species; while sulfite reduction is common among clostridia, it is not a universal phenotype for all toxigenic C. botulinum strains. Phenotypic expression is heavily influenced by strain-specific traits and assay conditions (Lindström and Korkeala, 2006; Peck, 2009). Consequently, while SRC-based methods are useful for assessing general hygiene, they have inherent limitations as predictors of C. botulinum (International Commission on Microbiological Specifications for Foods (ICMSF), 2014). These limitations may stem from high effective limits of detection (≤10 CFU/g) and inefficient spore germination or outgrowth on selective agar (Hauschild and Dodds, 1993).

While these constraints may be inconsequential in high-level contamination, they are critical at low levels observed here. Even low numbers of C. botulinum pose a significant risk to infants, who lack the established gut microbiota necessary to suppress colonization (Centers for Disease Control and Prevention (CDC), 1998; Lindström and Korkeala, 2006; Harris and Dabritz, 2024). The tiered analytical approach employed in this study – combining anaerobic enrichment with molecular screening – enabled the identification of samples requiring further investigation and guided targeted isolation efforts. Real-time PCR provided a sensitive screening tool, while amplicon sequencing and WGS provided the specificity and resolution necessary to link the outbreak definitively to the product (ISO/TS 17919:2013, 2013; Lindström and Korkeala, 2006; U.S. Food and Drug Administration (FDA), 2025).

4.1 Implications for industry

At the time of writing, the root cause investigation remains ongoing to determine likely sources of C. botulinum entry into the supply chain. The 2025 infant botulism outbreak – with 51 cases reported by the CDC as of publication – represents the first outbreak linked to PIF in which C. botulinum was detected in unopened finished product containers. This outbreak fundamentally changes the standing assumptions in PIF manufacturing: that C. botulinum is not an organism of concern requiring specific preventative controls.

The most significant consequence of this outbreak for the PIF industry is that C. botulinum can no longer be classified an unforeseen hazard. This necessitates an immediate reassessment of hazard analysis across the industry. Furthermore, these results highlight a critical consideration for industry testing programs: enumeration of SRC should be interpreted strictly as an indicator of general anaerobic spore presence rather than a reliable surrogate for C. botulinum detection. Reliance on SRC results alone can lead to false-negative conclusions; the test lacks the necessary sensitivity or specificity to serve as a proxy for toxigenic C. botulinum.

Incorporation of targeted molecular screening for botulinum neurotoxin genes – particularly within a tiered testing framework for high-risk products like PIF – can significantly enhance detection sensitivity while maintaining alignment with modern preventive control principles. Such an approach supports earlier detection, more informed risk-management decisions, and improved confidence in food safety assessments involving C. botulinum.

References

• 1

  American Public Health Association (2015). Compendium of methods for the microbiological examination of foods. 5th Edn. Washington (DC): APHA Press.

  • Google Scholar

• 2

  Centers for Disease Control and Prevention (CDC) (1998). Botulism in the United States, 1899–1996: handbook for epidemiologists, clinicians, and laboratory workers. Atlanta, GA: CDC.

  • Google Scholar

• 3

  Codex Alimentarius Commission. Food and Agriculture Organization of the United Nations. (2008). RCP 66 - code of hygienic practice for powdered formulae for infants and young children. Available online at: https://www.fao.org/input/download/standards/11026/CXP_066e.pdf (Accessed December 22, 2025)

  • Google Scholar

• 4

  DoyleM. P.BuchananR. L. (2013). Food microbiology: fundamentals and frontiers. 4th Edn. Washington (DC): ASM Press.

  • Google Scholar

• 5

  HarrisR. A.DabritzH. A. (2024). Infant botulism: in search of Clostridium botulinum spores. Curr. Microbiol.81:306. doi: 10.1007/s00284-024-03828-0,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 6

  HauschildA. H. W.DoddsK. L. (1993). Clostridium botulinum: ecology and control in foods. Food Microbiol.10, 255–269. doi: 10.1201/9781315139623

  • CrossRef
  • Google Scholar

• 7

  International Commission on Microbiological Specifications for Foods (ICMSF). (2014). Usefulness of testing for Clostridium botulinum in powdered infant formula and dairy-based ingredients for infant formula. Available online at: https://www.icmsf.org/wp-content/uploads/2018/02/ICMSF_Infant_Formula_Testing_Revision1-20140117.pdf (Accessed December 22, 2025)

  • Google Scholar

• 8

  ISO 15213-1:2023 (2023). Microbiology of the food chain—horizontal method for the detection and enumeration of Clostridium spp.—part 1: enumeration of sulfite-reducing Clostridium spp. by colony-count technique. Geneva: International Organization for Standardization.

  • Google Scholar

• 9

  ISO/TS 17919:2013 (2013). Microbiology of food and animal feed—polymerase chain reaction (PCR) for the detection of food-borne pathogens—detection of Clostridium botulinum. Geneva: International Organization for Standardization.

  • Google Scholar

• 10

  Joint FAO/WHO Expert Meetings on Microbiological Risk Assessment (JEMRA). World Health Organization. (2004). Enterobacter sakazakii and other microorganisms in powdered infant formula: meeting report. Available online at: https://www.who.int/publications/i/item/9789241562775 (Accessed December 22, 2025)

  • Google Scholar

• 11

  LindströmM.KorkealaH. (2006). Laboratory diagnostics of botulism. Clin. Microbiol. Rev.19, 298–314. doi: 10.1128/CMR.19.2.298-314.2006,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 12

  PeckM. W. (2009). Biology and genomic analysis of Clostridium botulinum. Adv. Microb. Physiol.55, 183–265. doi: 10.1016/S0065-2911(09)05503-9

  • CrossRef
  • Google Scholar

• 13

  PopoffM. R.BouvetP. (2013). Genetic characteristics of toxigenic Clostridia and toxin gene evolution. Toxicon75, 63–89. doi: 10.1016/j.toxicon.2013.05.003,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 14

  U.S. Food and Drug Administration (FDA) (2025). Bacteriological analytical manual (BAM), chapter 17: Clostridium botulinum. Silver Spring (MD): FDA.

  • Google Scholar