Botulism is a neurological disease common to humans and animals, caused by the action of botulinum toxins (BoNT) produced by bacteria from the genus Clostridium. There are seven BoNTs described historically, identified from A to G. Human botulism is mainly associated with toxins A, B, E and F (1) and animal botulism with toxins C, D and the mosaic forms C/D and D/C (1, 2). BoNT are recombinant BoNT types. BoNT C/D is composed of the light chain of BoNT C and the heavy chain of BoNT D and BoNT D/C is composed of the light chain BoNT D and the heavy chain of BoNT C (3). Botulism occurs on all continents and is variable in incidence. In all species, the disease presents with flaccid paralysis, including respiratory and heart failures (1). Animal botulism affects many species, mainly birds and cattle in France (4), but also fur animals (i.e., minks or foxes) in northern European countries (5, 6) and horses in the United States (7). Based on the current knowledge available, intoxication is the main mode of contamination of cattle at the origin of clinical signs. It is therefore the ingestion of preformed toxins in food, water or any contaminated substance that is currently considered the cause of botulism. Avian botulism is the result of consumption of Clostridium botulinum spores. It is assumed that toxin production occurs in vivo. Ingested spores germinate, proliferate and produce toxin primarily in the cecum. Absorption of toxin formed in the digestive tract is responsible for the symptoms (8). In humans, it is a rare disease. Five types of botulism are typically described in humans, depending on the mode of contamination and exposure to the toxin: foodborne botulism, intestinal botulism, wound botulism, iatrogenic botulism and inhalational botulism (9, 10). Foodborne botulism and infant intestinal botulism are the two most common forms observed (11).

Animal botulism is considered an emerging problem in Europe (12). At the European level, botulism is monitored through the surveillance of zoonoses and zoonotic agents and the protection of workers (exposure to biological agents at work). In France, the regulatory framework requires mandatory official notification, both in humans and in animals, regardless of the species affected. Human botulism has been monitored by the French health authorities since the establishment of the National Reference Center for Anaerobic Bacteria and Botulism (NRC, Institut Pasteur de Paris) in 1978 and reporting the disease to Santé Publique France (SPF) has been compulsory since 1986. Any suspicion of human botulism requires notification to the regional health agency (ARS) and its biological confirmation by the NRC. In animals, botulism has been regulated since 2006, first in poultry and then in wild birds and cattle. Until then, it was classified as a first category health hazard for all susceptible species (13). With the promulgation of the Animal Health Law at European level in 2016 (14), the status of this disease has changed, because it does not appear as such in the list of diseases transmissible to animals or humans that must be subject to fixed prevention and control measures. A National Reference Laboratory for avian botulism was designated in France in 2011 (NRL, ANSES Ploufragan-Plouzané-Niort Laboratory).

Case reports of human and animal botulism are regularly published, but studies compiling surveillance data on botulism are scarce, particularly with respect to animal botulism.

Here, we present the results of human and animal botulism surveillance based on SPF data as well as NRC and NRL biological investigations. First, annual variability in the occurence of botulism is discussed, followed by a description of the outbreaks observed.

The analysis of surveillance data made it possible to assess human and animal botulism in France. Our study confirms that the disease is present in many species, being rare in humans with an occurence of 10 persons affected per year, and much more common in animal species, essentially in birds (wild and poultry) and cattle, which are the two most affected categories of animals. Each year, on average, 10 outbreaks are recorded in the bovine sector, 30 in poultry sector and 20 cases in wild birds, each of which can affect several thousand birds (30).

At the European level, human data come from systems equivalent to the French mandatory reporting system (36, 37). These surveillance systems can be considered as exhaustive for the detection of severe forms of botulism. The number of confirmed cases over the 2011–2018 period was relatively stable with ~100 cases reported per year. The incidence rate in Europe is around 0.02 cases per 100,000 inhabitants, similar to the incidence rate in France (36). The countries with the highest number of confirmed cases are Italy, the United Kingdom, Poland, Romania and France. In Italy, 466 cases of botulism were identified from 1986 to 2015: 93% were foodborne botulism, infant botulism accounted for only 6% of cases and wound botulism for 1% (38). In Turkey between 1983 and 2017, 95 cases of botulism were identified, and the food category primarily responsible for the cases was home-canned vegetables (39). In Ukraine, between 1955 and 2018, 8614 cases of botulism were reported (40).

Infant botulism is the most common form of botulism in the United States and has accounted for 80% of reported cases of childhood botulism worldwide since this form of the disease was first recognized in 1976 (41). It has an average annual incidence rate of 2.1 cases per 100,000 live births (42). A recent review covering the 1976–2016 period identified 1345 cases (6.5 cases/100,000 live births/year) caused by types A, B, Ba, Bf and F in the state of California (43). The average annual incidence rate was calculated at 4.3 cases per million live births in Canada during the period 1979–2019 (44).

For animals, few data of surveillance are available at the global level and most of them come from France, where this disease is particularly monitored in animals. Botulism has previously been reported in 264 bird species representing 39 families (30). Anatidae is one of the most affected families, at least in France, as highlighted in our study and in at least one other study (16). In poultry, the species affected by botulism outbreaks are broilers, turkeys, pheasants and, to a lesser extent, ducks, laying hens (raised on litter or free-range only), geese, quails and guinea fowl (2, 16, 45–47). For cattle, only case reports are available in the literature and prevalence has not been reported.

Regarding other animal species, few are affected by botulism in France. A few cases were observed in domestic carnivores (cats and dogs) and only one case was reported in fish during this period. Information on the presence in fish is of great importance, because fish may be naturally affected by type E botulism responsible for human botulism. Mortalities due to botulism type E have been described around the world in wild species (e.g., the round goby Neogobius melanostomas in the Great Lakes region of North America, the catfish Ictalurus punctatus in the Mississippi Delta in the United States) (48–50). Regularly described on aquaculture farms from the 1960s to the 2000s (especially on trout or salmon farms), botulism outbreaks in aquaculture seem to have become rare, due in part to changes in farming and health management practices. The only relatively recent references, apart from those relating to cases of botulism E affecting fish in the Great Lakes region (48), involve botulism in catfish reported from some farms in the southern United States of North America (51, 52).

Our analysis of occurrences during the period studied here shows that the incidence of human botulism has been relatively stable over time. Similarly, animal botulism also appears to experience relative stability, although there are annual variations for which the origin cannot always be identified. Comparison over a longer period is made difficult by the changes in the animal botulism surveillance system in France over time and especially the significant development of diagnostic methods. Before 2010, the BoNT detection method did not allow the identification of mosaic forms, and different analytical methods were used to differentiate between BoNT types. The characteristics of the tests have also improved, with the optimization of sampling methods (choice of matrices, sampling protocol, transport and storage methods) (2, 46) probably leading to better sensitivity in regards to detection or diagnostic confirmation. Nevertheless, there are still situations, particularly in the bovine sector, where clinical suspicions strongly suggestive of the disease cannot be confirmed by laboratory analyses. For example, sera collected on symptomatic animals are often negative for BoNT using the mouse bioassay (45), probably because BoNT is not circulating any more when the sample is collected. A difference in sensitivity to BoNTs between mice and cattle could also be hypothesized. It has indeed been suggested that cattle are 12.88 times more sensitive to BoNT C than a mouse on a per kilogram weight basis (53), BoNT C has moreover been shown to be the least toxic BoNT types for mice (54). On the contrary, mice are extremely sensitive to BoNT D/C, which harbored the highest toxic activity among tested BoNT types (54). While a difference in sensitivity between mice and cattle may explain the failure of the mouse bioassay to detect BoNT C in serum samples from cattle, this seems unlikely as far as BoNT D/C concerns considering the high sensitivity of mice to this BoNT type. Detection of BoNT-producing clostridia could be sometimes tricky when contamination is low and not homogenous within the matrix. Several matrices (liver, ruminal content, fecal samples…) collected from different symptomatic animals should be analyzed to make sure C. botulinum will be detected.

The cases presented in this report correspond to those identified by the NRC and the NRL. In France, any suspicion that is submitted for laboratory diagnosis currently goes through a reference laboratory. The severe forms of human botulism are probably reported exhaustively (mild forms may not be detected in humans, e.g., solely involving digestive discomfort), but it is likely that a certain number of animal botulism suspicions are not reported, and their extent cannot be assessed. This under-reporting is probably limited in the cattle sector. In the poultry sector, because botulism outbreaks occur at the end of the rearing period, we cannot exclude the possibility of flocks being sent to slaughter at the start of an outbreak of botulism. Surveillance of botulism in wild birds, which is based on event-based surveillance, leads to an obvious under-representation of cases, but it is not possible to assess to what extent.

Analysis of the toxin types occurring in France confirmed the predominance of types A and B in human botulism—in both foodborne and infantile cases—and exceptionally type F (55, 56). At the international level, BoNT types that cause human cases are types A and B, followed by E and, occasionally, F. A meta-analysis of outbreaks including 197 outbreaks of foodborne botulism (nearly half of which involved outbreaks in the US) identified BoNT A, B, E, and F as the causative BoNT in 34, 16, 17, and 1% of outbreaks, respectively (57). BoNT B is the most prevalent BoNT in France, like in Poland where type B represented 83% of the cases in 2016 (58). In Italy, from 1986 to 2015, BoNT B was involved in 79.1% of cases (261/330), followed by BoNT A (9.7%, 32/330), with BoNT F, Ab, and Bf, accounting for 0.3 (1/330), 1.5 (5/330), and 0.6% (2/330) of all cases, respectively (38). In Ukraine, BoNT B (59.64%), E (25.47%), and A (7.97%) are the most common, with cases related to BoNT C being very minor (0.56%) and only suspected (40). In North America, foodborne botulism outbreaks originate from vegetables (home-canned), but mostly BoNT E, originating from fish or marine mammals prepared in indigenous communities using traditional methods (e.g., fish fermentation) (42). Similarly, in various Asian countries, outbreaks typically arise from traditional food preparations (59–61).

The C/D mosaic form is the predominant BoNT in birds in France. BoNT C and D are also observed, but to a lesser extent. Other European countries report similar findings on field collections of strains from animal botulism outbreaks (3, 62). Although the majority of bird species are experimentally sensitive to various BoNTs, the only BoNTs naturally involved in outbreaks in birds are BoNTs C, D or their mosaics C/D and D/C, BoNT E and, much more rarely, BoNT A (63). BoNTs C, D or their mosaics C/D and D/C are the most frequent, both in wild and domestic species. BoNT E is less frequently detected, and regularly causes sporadic cases or epizootics in wild fish-eating birds in northern regions, but is rarely the cause of epizootics in farmed species. Type A botulism has only been described a few times in the United States in avifauna including deaths of seagulls in the Klamath River basin in California (63) and it seems to be excessively rare on farms [one outbreak on a broiler farm in the United States, see Graham and Schwarze (64)].

In France, only BoNTs D/C and C have been identified in recent years in bovine botulism outbreaks. In Europe, BoNT D/C is the currently cause of the majority of bovine cases (62, 65). Very rare cases with BoNT A were reported in the middle of the 20th century in France Prévot et al. (66, 67) cited by the French Agency for Food safety and Animal health (68), in zebus (Bos indicus) in Brazil (69), in dairy cows in Egypt in 1976 (4) and very recently in the state of New-York in the United States (70). Type B outbreaks have also been described in the literature in dairy herds: in the United States in 1984, 1992 and 2001, in Israel in 2000 (71–74) and in the Netherlands in about 30 dairy herds in 1976 and 1977 in the Netherlands in connection with the incorporation of contaminated brewers' grains in the feed ration (75).

The detailed analysis conducted on NRL data on avian botulism provided interesting details, particularly for poultry farms (species involved, age of onset of cases, dominant toxin types by species, etc.). In our data, there were no differences between males and females in poultry farming. In the literature, males appear to be more affected than females, particularly in turkeys (47, 76, 77). No explanation for this observation has been provided to date. For example, males have a longer rearing period than females, but the impact of this factor on the occurrence of an outbreak of botulism has not been evaluated. Botulism can also occur as a result of stress or a biosecurity failure at the time of removal of the females. Most of the outbreaks in both cattle and poultry are located in Brittany, in an area with a high density of poultry and dairy farms, and a high number of mixed farming, which may explain frequent cross-contamination and this higher prevalence. It is unlikely that there is detection bias, because the level of disease surveillance is the same throughout the country.

It was not possible to conduct a detailed analysis on cattle due to the lack of previous data, available only since 2017. In any case, this analysis remains difficult to conduct retrospectively. The information available is based almost exclusively on more or less complete information forms accompanying the samples sent to the NRL. A standard information form listing the essential data to be transmitted with the samples would facilitate the monitoring of animal botulism in France.

Addressing the study of pathogens not sector by sector but from a global perspective is the basis of the One Health concept. This approach address a health threat at the human-animal-environment interface based on collaboration, communication, and coordination across all relevant sectors and disciplines with the ultimate goal of achieving optimal health outcomes for both people and animals (78). Botulism is part of the European list B of Annex I of the zoonoses Directive (79), surveillance and study of botulism, BoNTs, and BoNT-producing clostridia logically fall under the One Health concept. If botulism is notifiable for humans in Europe, this is not systematically the case for animals. In France, botulism is a notifiable disease, both in humans and animals, regardless of the species affected, which allows for an overall view. The occurrence of botulism cases and outbreaks is closely monitored through a case-based, passive surveillance system. This is a first step in the application of the One Health approach to disease surveillance by juxtaposing animal and human surveillance. In the majority of cases, surveillance systems continue to be developed and operated within a highly sectoral approach (80). But to be effective, the management of complex health issues should shift from isolated, sectoral and linear, to systemic and transdisciplinary approaches to health (81). Our study has shown that human botulism is mostly due to ham (pig sector) and canned vegetables, indicating the importance of collection of surveillance data from food industry, animal sectors as well as surveillance of this pathogen in the environment. These results show that even if surveillance is implemented for both human and animal health, progress are still needed to improve data collection and surveillance of food, feed sectors and environmental contamination.

With the exception of BoNT E, which was exceptionally detected throughout the study period in wild birds, the types of botulism found in animal outbreaks are different from those identified in human outbreaks over the last 10 years in France and no human botulism outbreaks investigated by SPF and the NRC have been linked to animal botulism. But both human and animals are known to be sensitive to some similar BoNT types. As a result, detecting a BoNT E outbreak in wild birds or in poultry, or a BoNT B outbreak in cattle is crucial to prevent any contamination to humans. Furthermore, there are currently very few cases of type C, D, C/D, D/C in humans. It is important to continue to monitor over time that this is still the case. Early detection of zoonotic pathogens through enhanced laboratory capacity and surveillance at the animal–human interface is a crucial step toward controlling and preventing zoonoses (82).

Given that botulism is ubiquitous in the environment and can cause disease in both humans and animals, it is essential to enhance links between human and animal surveillance systems. Accordingly, in line with the One Health concept, this study presents the first integrative approach to the routine surveillance of botulism in humans and animals in France.

The original contributions presented in the study are included in the article/Supplementary material, further inquiries can be directed to the corresponding author.

SL, CLM, RS, LG, and CM were responsible for the concept and design of the study, interpretation of results, writing, and critical review of the manuscript. CLM, RS, and LG were responsible for data collection. CL and LG were responsible for statistical analysis. CLM, KP, PK, and FM were responsible for reviewing and editing the manuscript. All authors contributed to the article and approved the submitted version.

This work was supported by the French Ministry of Agriculture and Food.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.