Formal and informal antimicrobial trade and usage in farmed animals of the urban area of Lomé, Togo

Abstract

Inappropriate antibiotic use in livestock is a key driver of antimicrobial resistance. Surveillance and prevention efforts in low-income countries are challenged by widespread informal sales and administration. We investigated the antibiotics trade network for farm animals and farmers' antibiotic purchasing and usage behaviors in an urban region with high human and animal population density and prevalence of antimicrobial resistance.

We conducted semi-structured interviews with 116 farmers in Lomé, Togo, in areas where multidrug-resistant bacteria had been identified. Using multiple-factor analysis and hierarchical clustering, farmers were categorized based on their livestock production, antibiotic purchase and usage patterns, and alternative disease management strategies. Additionally, 44 antibiotic distributors were interviewed using snowball sampling, allowing us to reconstruct the distribution network and its key actors.

Commercial poultry and dog farmers exhibited significantly higher antibiotic consumption per unit of livestock biomass than other farmers, seeking to secure their production. Smallholders used antibiotics less frequently, often replacing them with phytotherapy to minimize costs. Large-scale farmers sourced antibiotics from legal distributors, whereas smallholders combined legal and illegal sources. Market wholesalers and retailers, mostly women, were central to the illegal antibiotic trade, primarily supplied through cross-border smuggling.

Our findings highlight the need for antimicrobial resistance prevention programs that consider farmers' specific antibiotic use behaviors and motivations. Addressing the illegal antibiotic trade requires a gender-sensitive approach, as women dominate informal distribution networks. Understanding these dynamics is crucial to designing effective interventions to curb antibiotic misuse and mitigate risks of antimicrobial resistance.

Highlights

• •A quantitative analysis of the antibiotics distribution networks supplying farms.
• •Trade in and use of antibiotics on livestock farms in the city of Lome and its suburbs.
• •Gender roles among antibiotic users and actors in the distribution system.

1. Introduction

Since the discovery of penicillin in 1928, the use of antibiotics (ABs) to treat or prevent bacterial infections has become essential to human and veterinary medicine, drastically reducing the morbidity and mortality of many infectious diseases in humans and animals [1]. According to Klein et al., between 2000 and 2015 the global consumption of ABs increased by 65 % globally and by 114 % specifically in low- and middle-income countries [2]. However, the effectiveness of ABs is being undermined by the development of antimicrobial resistance (AMR) in pathogenic bacteria, ultimately rendering AB-based treatments ineffective against bacterial infections [3]. The number of deaths directly linked to AMR was estimated to be more than 1.2 million in 2019, and this figure is expected to increase to around 10 million deaths a year by 2050 [4]. The development of AMR in human health is a phenomenon with multiple causes, including excessive and inappropriate antimicrobial use (AMU) [5]. A growing number of people are self-medicating, using ABs without the consultation of a health professional [5,6]. The over-the-counter sale of ABs, both in pharmacies and on informal markets, enables people to acquire these medicines without a prescription, increasing the risk of AB usage with inappropriate doses or treatment duration or contraindications [7]. Poor prescription practices, which may stem from both prescriber and consumer behavior, lead to inappropriate use of AB. The use of counterfeit or out-of-date products available on the market is an additional aggravating factor [5,6]. In Africa, the lack of control of the access to ABs and the inadequate management of out-of-date medicines exacerbate this phenomenon. Since the 1980s, this has been compounded by the development of a parallel market in ABs for both human and veterinary use, sold outside official channels [5,8].

Recent results confirm the role of the use of ABs in livestock farming in the development of AMR in human health. An analysis of the phylogenetic distance between samples of Escherichia coli isolated in samples from humans, animals, and the environment in a rural area of Madagascar identified 104 cases of transmission of Escherichia coli carrying beta-lactamases (ESBL) between compartments [9]. Another study carried out in the Netherlands revealed that 19 % of multi-resistant bacteria (MRB) observed in humans came from food of animal origin, while 60 % were attributed to human-to-human transmission [10]. These findings underline the importance of adopting a “One Health” approach to the surveillance and management of AMR, as recommended in the global action plan of the World Health Organization (WHO) [11]. The misuse of ABs in farm animals has been documented in several studies. Farmers routinely re-use ABs to treat several animals with similar symptoms, without receiving the advice of an animal health expert [5,12] or administer ABs for non-therapeutic purposes, such as growth promotion in poultry. This further increases the risk of development of resistance due to prolonged exposure to the animals' bacterial flora to ABs [12]. In Vietnam, a study conducted between 2016 and 2017 revealed that more than 50 % of farmers perceived ABs as feed supplements intended to improve the general condition of animals, and not specifically intended to treat bacterial infections. According to the same study, 59.4 % of farmers used ABs for non-therapeutic purposes [13]. Another study carried out in Kenya in 2024 showed that 58 % of farmers administer ABs themselves without medical advice, although 76 % said they were aware of the risk of AMR development [12].

Another major factor contributing to AMR is the supply of low-quality ABs to farmers, containing inadequate concentrations of active substances, often obtained in unregulated markets [14]. Clifford et al. advocate for a better monitoring of the informal medicine supply chains in low- and middle-income countries, where many farmers resort to them, because of financial constraints or limited access to quality medicines, but this requires a thorough understanding of their structure [8].

Sub-Saharan Africa is considered to be one of the regions most affected by resistance to ABs in the world, with around 250,000 deaths a year attributed to treatment failure due to the ineffectiveness of ABs, representing a mortality rate of 27.3 deaths per 100,000 inhabitants, compared to 6.5 deaths per 100,000 in Australia [15]. In some West African countries, this rate even exceeds 100 deaths per 100,000 inhabitants [16]. Several studies have reported sub-optimal use of ABs in livestock farming, which varies according to the context and production system. In a survey conducted in five East African countries, Caudell et al. (2020) found that 70 % of poultry farmers had a correct understanding of the use of ABs, compared with 40 % of pastoral ruminant farmers [17]. In Tanzania, Maasai herders and nearby agropastoralists in the Arusha region had the highest frequency of administration of ABs by non-professionals (over 90 % and 70 % of respondents respectively), compared with farmers in the Chagga zone (less than 5 %) [18]. In Ghana, a study conducted by Boamah and Agyare (2016) on poultry farms revealed that the use of ABs is positively correlated with farm size, the presence of other animal species on site, as well as the presence of particular infections such as coccidiosis and chronic respiratory diseases [19].

Research into the use of ABs is still limited in West Africa, particularly in the Gulf of Guinea region (Ghana, Togo, Benin, Nigeria) [17]. Although this area is not yet considered to be at elevated risk for antibiotic resistance in animals [20] it is characterized by a high human population density associated with large urban centers. Additionally, the intensive use of ABs in livestock farms in this region represents a well-known risk factor for AMR. In Togo, a study revealed that out of 1377 bacterial strains isolated in human patients at hospital, 309 (22.44 %) were resistant to third-generation cephalosporins [21]. The use of ABs and subsequent development of AMRs in domestic animals sharing the same environment with humans is one of the potential explanations for the high rate of MRB carriage in the human population. However, our limited understanding of the use and distribution of ABs in the living areas of MRB carriers compromises the promotion and implementation of good AMU practices, adequate control measures of AB supply, and effective surveillance of AMR [21,22].

The aim of this study was to characterize AB usage, ABs perception, and ABs supply chain in domestic animals in the living areas of MRB carriers of the Lomé metropolitan area, capital city of Togo and one of the major urban centers in the Gulf of Guinea region. More specifically, the study aims to (1) characterize the distribution of AB use practices among the domestic animal farmers by identifying clusters of farmers sharing similar socioeconomic features, farming system characteristics and practices related to AB use, (2) describe and analyze the distribution circuit for ABs intended for livestock, identifying the formal and informal actors involved in the distribution and their relative importance in the circuit, and (3) describe the knowledge and perception of livestock farmers and actors of AB distribution towards the use of ABs and alternatives to ABs for animal health management. Our study encompassed all classes of ABs (i.e. any antimicrobials used against bacterial infections).

2. Materials and methods

2.1. Target population and sampling method

A socioeconomic survey was carried out in the maritime region of Togo, which encompasses the urban area of Lomé. The target population of the study were the domestic animal farmers (hereafter referred to as “farmers”) located in the living area of human patients identified as MRB carriers, as well as people involved in the supply of AB used in these farms. Domestic animal farms were defined as operations aimed at breeding and/or keeping domestic animals for the purpose of sale or consumption of live animals or their products. It included included both livestock farms (e.g. pigs, cattle, goats, sheep, poultry, or rabbits) and companion animal farms (e.g. dogs).

The identification of farmers was performed in two successive steps. Patients treated in the traumatology-orthopedics, intensive care, pediatrics, and gynecology departments of Lomé's two university hospitals were identified and microbiological analyses were carried out on stool samples taken from these patients, in order to detect a carriage of MRB, defined as either Extended-spectrum β-lactamase (ESBL)-producing Enterobacteriaceae or Carbapenemase-producing Enterobacteriaceae (CPE). Details on the selection of hospital patients and performed microbiological analysis and their results are provided in Appendix I. The houses of patients with confirmed MRB carriage were visited. If a patient happened to be a domestic animal farmer, he was included in the socioeconomic study. Additionally, domestic animal farms located less than 2 km away from patients' houses were searched for and included in the study through a convenient sampling procedure. The farms were searched within a 2 km radius around the houses of the MRB carriers until three farms were identified. In some instances, however, only two farms domestic animal farms could be found. We chose to limit the investigations to 3 farms at maximum per hospital patient in order to balance the objective of maximizing the geographical coverage of the study with the logistical constraints that limited the number of feasible visits.

During their interviews, farmers were asked to describe their use of AB on animals, and list the distributors, formal or informal, from whom they obtain ABs. To reconstruct the distribution chain a snowball sampling approach was used, consisting in using the information gathered during initial interviews to identify and include new participants. The mentioned distributors were contacted and invited to take part in an interview and, in turn, listed the people from whom they purchased ABs, who in turn listed their AB suppliers. Following this procedure, the distribution network, from initial suppliers down to the final end users (i.e. the farmers) was reconstructed. All interviews, including farmers and AB distributors, were conducted from February 2023 to March 2024.

2.2. Structured and semi-structured interviews

Individual semi-structured interviews were used with farmers and distributors to gather information on farm management, use of ABs, ABs distribution, and perception of issues related to animal health management and AB usage. The farmer interview guide consisted of two sections: the first contained personal information to identify the farmer, and the second listed different themes that were addressed successively: (1) the description of the respondent's farm, (2) the use of ABs, (3) their origin, (4) sources of advice on health management and use of ABs, (5) possible measures to reduce the use of ABs, and (6) knowledge, use and perception of alternatives to ABs. This semi-structured interview was supplemented by a structured questionnaire interview covering specific aspects: (1) the animal species reared and the type of production, (2) the intended use of animals, and (3) the biosafety measures in place. For distributors, the themes covered during the semi-structured interview were (1) the respondent's professional experience and activity, (2) the veterinary products sold, (3) the feed supplied, (4) the origin and destination of the ABs, (5) the importance ABs in the participant's income, (6) knowledge of the problems associated with the use of ABs, and (7) knowledge of alternatives to the use of ABs. The evaluation of the relative importance of different items on the same issue (e.g. the relative proportion of ABs used by a stakeholder from different sources) was collected from participants (farmers or distributors) using the proportional pilling method commonly used in several studies, particularly in the field of participatory epidemiology [23]. It consisted of asking the respondent to distribute 50 beans across the various items represented by circles drawn on a piece of paper. The semi-structured interview guides for the distributors and the farmers and the questionnaire administered to the farmers are available in Appendix II, III, and IV respectively.

2.2.1. Data entry

The data obtained from the semi-structured interviews was first recorded in a notebook, and then entered into an electronic form developed in “Kobocollect” (https://kf.kobotoolbox.org). The data from the structured questionnaire interview was entered directly during the interview on another dedicated electronic form on “Kobocollect”. The interview guides and questionnaires were developed by the study investigators and tested in pilot phases with 5 farmers before the survey began. Following the pilot phases, all initial themes and questions were kept and no additional theme and question were added, but some of the questions of the structured questionnaire were reformulated to ease their understanding.

2.2.2. Data processing

The intensity of AB use of farmers was assessed by estimating the annual financial cost of AB purchase per unit of biomass of animals kept in the farm, measured in tropical livestock unit (TLU). An approximation of the animal population biomass of the farm in TLU was used, instead of the livestock herd size, in order to harmonize and facilitate comparison of sizes between farms having different species composition. The cost was expressed in local currency, the « Franc de la Communauté Financière d'Afrique de l'Ouest » (CFA). The biomass of each animal species present on the farm was calculated by multiplying the number of animals of each species by the unit biomass of the corresponding species (expressed in TLU) [24]. The annual purchase cost of ABs per farm was divided by the sum of the amount of biomasses of all the species present on the farm, using the following formula:

$$\mathit{cost}\mathit{AB}\mathit{per}\mathit{TLU}=\frac{\mathit{annual}\mathit{cost}\mathit{AB}}{\sum \mathit{animal}\mathit{species}\mathit{biomass}\left(\mathit{TLU}\right)}$$

And the biomass of a given farm species was:

$$\mathit{animal\; species\; biomass}\left(\mathit{TLU}\right)=\mathit{Number\; of\; animals\; of\; the\; species}\times \mathit{TLU}\mathit{species\; coefficient}$$

2.3. Data analysis

The first objective was to characterize the distribution of AB use practices among the farmers by identifying clusters of farmers sharing similar socioeconomic features, farming system characteristics and AB use practices. To this end, we established a typology of the interviewed farmers based on a set of variables collected during questionnaire and semi-structured interviews. A basic description of the data was carried out using simple metrics. A multiple factor analysis was carried out to reduce the dimensionality of the variables in order to group the surveyed farmers into clusters (i.e. groups of farmers) according to the characteristics of the respondents and their farms, including socio-economic and farm management variables, variables related to animal health management and AB use and farmers' knowledge and use of alternatives. This approach allowed us to: (1) analyze datasets with both quantitative and qualitative variables; (2) group variables into separate themes and balance the contribution of each theme (with different numbers of variables) by weighing the variables included in each theme so as to equalize their maximum influence (inertia), thus preventing any theme from having a disproportionate influence on the results [25]. The method locally combines the principles of principal component analysis for quantitative variables and those of multiple correspondence analysis for qualitative variables [25]. We grouped the quantitative and qualitative variables into the following themes: (1) characteristics of the farmer and his farm (qualitative variables), (2) characteristics of the farmer (quantitative variables), (3) farm management, including biosecurity (qualitative variables), (4) use of ABs (qualitative variables), (5) use of ABs (quantitative variables), (6) knowledge and use of alternatives to antibiotics (qualitative variables). The farmers were clustered using the Hierarchical Clustering on Principal Components method. A hierarchical tree was built and the optimal number of clusters was determined by maximizing the loss of relative inertia resulting from successive partitions of the tree [26].

The second objective was to describe and analyze the distribution circuit for ABs intended for livestock. A basic description of the AB distributor data was carried out to define and characterize the different categories of distributors surveyed. The AB supply network, including the different categories of distributors and the flow of ABs between these categories, was reconstructed. The supply network was strictly hierarchical, with categories of distributors located on 4 distinct levels and a level 0 being the AB end-users (the farmers). All identified flows of ABs were directed from one category to another situated on a strictly lower level. For each category of distributor, i we identified an ensemble $P\left(i\right)$ of paths p connecting i to farmers (level 0) through direct AB sale to the farmers, or, indirectly, through successive transactions with intermediaries noted $j\in \left\{01,\dots n\left(p\right)\right\}$, 0 corresponding to farmers and $n\left(p\right)$ the number of categories of distributors on the path p. It was, therefore, possible to estimate the fraction ${\mathrm{Y}}_i$ of the value of ABs purchased by farmers transiting through each category i using the following equation:

$${\mathrm{Y}}_i={\sum }_{p\in P\left(i\right)}{\prod }_{j=0}^{n\left(p\right)}{x}_{j,j+1}$$

With $x_{j,j+1}$ the proportion of the ABs purchased by the distributors of category j coming from distributors of category j + 1. The rational is illustrated in Appendix V. For any category of distributor j, $x_{j,j+1}$ was estimated with the following equation:

$$x_{j,j+1}=\frac{{\sum }_{d\in D_j}{w_{j+1}}_d}{n_j}$$

$D_j$ being the ensemble of interviewed distributors of category j, $n_j$ the number of interviewed distributors in category j (i.e. the size of $D_j$), d any distributor belonging to $D_j$, and ${w_{j+1}}_d$ the proportion of ABs purchased by distributor d from the distributors of category j + 1, estimated using the proportional piling tool. In order to take into consideration, the differences in quantities of ABs purchased across individual farmers, the average proportion of ABs purchased by farmers (level 0) from the distributors of category 1 ($x_{0,1}$) was estimated as a weighted average, using the individual annual ABs expenses as weights:

$$x_{0,1}=\frac{{\sum }_{f\in F}{c}_f{w_1}_f}{{\sum }_{f\in F}{c}_f}$$

F being the ensemble of farmers and f any farmer belonging to F, $c_f$ the annual expenses in AB purchase by farmer f and${w_1}_f$ the proportion of ABs purchased by f from distributors of category 1, estimated using the proportional piling tool.

The third objective was to describe the knowledge and perception of livestock farmers and actors of AB distribution towards the use of ABs and the alternatives to ABs for animal health management. We described the frequency of responses to the related questions in each of the identified cluster of farmers, and in the informal and formal actors of AB distribution.

2.4. Data processing and analysis tools

Data management and analysis were carried out using R software version 4.4.0 [27]. Data cleaning and management were carried out using the “dplyr” and “tidyr” packages, and multiple factor analysis was performed using the R packages “factoMineR” and “factoextra”. Graphical representations were produced using the “ggplot2” and “ggrepel” packages.

2.5. Ethical procedure: involvement of human participants and data protection

The data collected includes the personal data of human participants. It was collected, stored, processed, and shared in compliance with the European Union's General Data Protection Regulation (2016/679). The interviews were conditioned by the explicit approval of the participants, documented by signing an informed consent form. This form, which was explained orally to the participants, consists of several parts. The first part contains information essential for understanding the study and the processing of interview data: the name of the project, the study title, the members of the research team, the objective of the study, the addressed themes, the data protection procedures, and the participants' right to interrupt the survey and access, rectify or delete their personal data. The second part contains questions designed to check that the information provided has been understood, and to ensure that the participant is aware of his or her rights, in particular the right of withdrawal and control over his or her personal data. A copy of the informed consent form is available in Appendix VI and VII. Survey data was shared only between the researchers directly involved in data analysis. The study was evaluated and approved by the Research Ethics Committee of the Ministry of Health of Togo (opinion number 043/2022/CBRS).

3. Results

116 semi-structured interviews with farmers and 44 with distributors of ABs were conducted in the commune of Lomé and the surrounding area from February 2023 to March 2024. The spatial distribution of these farmers is displayed in Appendix VIII. Using multiple factor analysis, we identified six distinct clusters (Appendix IX comprising 54 (cluster 1), 30 (cluster 2), 7 (cluster 3), 19 (cluster 4), 5 (cluster 5), and one (cluster 6) farmers. Due to the small size of cluster 6, made up of a single farmer with an atypical profile (very large multi-species farm), we decided to focus our analysis on the first five clusters. In all the clusters, the vast majority of respondents own their livestock (Table 1), they all produce the livestock feed by themselves with cereals and plants purchased or grown on the farm, and there was no mention of the addition of ABs to the feed. Most of them mentioned financing difficulties as a major issue (Table 2). Regardless of the cluster, drug treatments were the most frequently cited animal diseases management method, with the majority of farmers declaring using ABs, and the most frequently cited molecule being oxytetracycline (Appendix X). The clusters, however, differ on several aspects detailed in the following paragraphs.

Table 1.

Distribution of farmer characteristics among the 5 clusters identified from the Multiple Factor Analysis carried out on the 116 farms surveyed in the urban area of Lomé between February 2023 and March 2024. For qualitative variables, we present the number and percentage of participants with the considered characteristic (e.g. female participants) within each cluster. For quantitative variables we present the mean value and the range (minimum – maximum) within each cluster.

Variables	Cluster 1 (n = 54)	Cluster 2 (n = 30)	Cluster 3 (n = 7)	Cluster 4 (n = 19)	Cluster 5 (n = 5)
Participants who own their farm	52 (96.3 %)	30 (100.0 %)	6 (85.7 %)	18 (94.7 %)	5 (100.0 %)
Female participants	19 (35.2 %)	15 (50.0 %)	0	1 (5.3 %)	0
Age of participants (years) (min-max)	53 (23–80)	44 (27–68)	38 (10–61)	53 (14–77)	44 (33–61)
Number of children (min-max)	5 (0–16)	3 (0–6)	4 (0–7)	7 (2–18)	2 (0–7)
Number of years of experience in farming (min- max)	12 (1–44)	9 (1−30)	9 (4–13)	20 (2–34)	10 (5–19)
Participants with a fridge at home	19 (35.2 %)	6 (20.0 %)	5 (71.4 %)	5 (26.3 %)	4 (80.0 %)
Participants considering farming as a main, secondary or negligible source of income
Main	5 (9.3 %)	3 (10.0 %)	3 (42.9 %)	14 (73.7 %)	5 (100.0 %)
Secondary	45 (83.3 %)	24 (80.0 %)	4 (57.1 %)	5 (26.3 %)	0
Negligible	4 (7.4 %)	3 (10.0 %)	0	0	0
Levels of education of participants
Not enrolled	13 (24.1 %)	7 (23.3 %)	0	5 (26.3 %)	0
Primary	15 (27.8 %)	10 (33.3 %)	1 (14.3 %)	5 (26.3 %)	0
Secondary	22 (40.7 %)	10 (33.3 %)	3 (42.9 %)	9 (47.4 %)	2 (40.0 %)
University	4 (7.4 %)	3 (10.0 %)	3 (42.9 %)	–	3 (60.0 %)

Table 2.

Distribution of farm characteristics among the 5 clusters of farmers identified from the Multiple Factor Analysis carried out on the 116 farms surveyed in the urban area of Lomé between February 2023 and March 2024. We present the number and percentage of participants with the considered characteristic (e.g. the participants who practice sanitary emptying) within each cluster.

Variables and variable groups	Cluster 1 (N = 54)	Cluster 2 (N = 30)	Cluster 3 (N = 7)	Cluster 4 (N = 19)	Cluster 5 (N = 5)
Number of buildings in the farm
0 buildings	2 (3.7 %)	0	0	0	0
1 building	46 (85.2 %)	25 (83.3 %)	4 (57.1 %)	12 (63.2 %)	0
2 buildings	4 (7.4 %)	2 (6.7 %)	3 (42.9 %)	7 (36.8 %)	3 (60.0 %)
≥ 2 buildings	2 (3.7 %)	3 (10.0 %)	0	0	2 (40.0 %)
Buildings' characteristics (closed or open)
Closed	21 (38.9 %)	13 (43.3 %)	6 (85.7 %)	7 (36.8 %)	5 (100.0 %)
Open	33 (61.1 %)	17 (56.7 %)	1 (14.3 %)	12 (63.2 %)	0
Participants who practice sanitary emptying	0	0	0	1 (5.3 %)	3 (60.0 %)
Difficulties encountered in animal farming by participants
Animal diseases	26 (48.1 %)	20 (66.7 %)	1 (14.3 %)	10 (52.6 %)	1 (20.0 %)
Animal theft	16 (29.6 %)	14 (46.7 %)	1 (14.3 %)	4 (21.1 %)	0
Animal mortality	2 (3.7 %)	2 (6.7 %)	0	2 (10.5 %)	0
High input prices	2 (3.7 %)	1 (3.3 %)	2 (28.6 %)	0	1 (20.0 %)
Low or unstable product prices	1 (1.9 %)	1 (3.3 %)	0	0	1 (20.0 %)
Financing difficulties	28 (51.9 %)	19 (63.3 %)	5 (71.4 %)	14 (73.7 %)	4 (80.0 %)
Difficulties in obtaining veterinary products	0	1 (3.3 %)	0	0	0
Difficulties in obtaining a veterinary service	2 (3.7 %)	0	0	1 (5.3 %)	1 (20.0 %)
Conflict with neighbors	5 (9.3 %)	2 (6.7 %)	0	3 (15.8 %)	1 (20.0 %)
Land-tenure issues	8 (14.8 %)	2 (6.7 %)	2 (28.6 %)	1 (5.3 %)	0
Other	6 (11.1 %)	6 (20.0 %)	3 (42.9 %)	5 (26.3 %)	2 (40.0 %)
Animal health management practices of participants
Drug treatments	29 (53.7 %)	25 (83.3 %)	5 (71.4 %)	13 (68.4 %)	4 (80.0 %)
Vaccination	8 (14.8 %)	0	2 (28.6 %)	6 (31.6 %)	2 (40.0 %)
Deworming	11 (20.4 %)	11 (36.7 %)	4 (57.1 %)	8 (42.1 %)	2 (40.0 %)
Administration of vitamins	6 (11.1 %)	9 (30.0 %)	5 (71.4 %)	7 (36.8 %)	2 (40.0 %)
Adequate power supply	5 (9.3 %)	2 (6.7 %)	1 (14.3 %)	1 (5.3 %)	1 (20.0 %)
Hygiene	3 (5.6 %)	1 (3.3 %)	4 (57.1 %)	3 (15.8 %)	1 (20.0 %)
Disinfection		1 (3.3 %)	0	1 (5.3 %)	0
Calling animal health professional to treat diseases	23 (42.6 %)	6 (20.0 %)	4 (57.1 %)	7 (36.8 %)	1 (20.0 %)
Phytotherapy	4 (7.4 %)	13 (43.3 %)	0	3 (15.8 %)	0
Other	2 (3.7 %)		0	0	0
No specific practice	4 (7.4 %)	0	0	1 (5.3 %)	0

3.1. Cluster 1: small-scale farmers, moderate informal users of ABs, unaware of alternatives

This cluster is composed of small-scale multi-species livestock farmers, with an average herd size of 2.1 TLU, frequently involved in goat farming (69 %, Fig. 1A, B), and for whom livestock farming is most often a secondary source of income (83.3 %, Table 1). A large proportion of these farmers are women (35.2 %), and most of them have had no education beyond primary school and do not own a fridge at home (Table 1). 68.5 % of the farmers in this cluster administer ABs, at an average cost of 40,481 FCFA/TLU/year, mainly to treat sick animals (59 % of farmers, Fig. 1D). The ABs used are more frequently obtained from informal sources (33.3 %) than from formal sources, such as veterinary pharmacists (22.2 %) (Fig. 2A). These farmers seek advice from a variety of sources, including livestock technicians (32 %), veterinarians (26 %), and livestock veterinary auxiliaries (LVA) (20 %), while a significant proportion rely on their own experience (28 %, Fig. 2B). Farmers in this cluster show limited knowledge and adoption of alternatives to ABs (17 % and 9 %, respectively). Their use is concentrated exclusively on phytotherapy, motivated by its efficacy (6 %) and the high cost of antibiotics (4 %, Fig. 3A, B and C).

Fig. 1.

Characteristics of livestock and livestock farmers in the five clusters identified on the basis of survey data from the urban area of Lomé (February 2023–March 2024): animal species present (A); Total livestock biomass (TLU = tropical livestock unit) (B); Antibiotic usage and expenditures: Frequency of reasons for antibiotic usage by farmers (C) and boxplot representation of the cost of antibiotics per unit of biomass (D) in the five clusters of farmers identified on the basis of survey data from the urban area of Lomé (February 2023–March 2024).

Fig. 2.

Antibiotic and advice sources by cluster: Frequency of sources of antibiotics (A) and sources of advice on health management (B) among livestock farmers in the five clusters identified on the basis of survey data from the urban area of Lomé (February 2023–March 2024).

Fig. 3.

Alternatives: Alternatives to antibiotics known (A) and used (B) by farmers and reasons (C) and obstacles (D) to them in the five clusters of farms identified with the survey data from the urban area of Lomé (February 2023–March 2024). The number of farmers aware of the existence of alternatives to antibiotics is 9 (16.7 %) in cluster 1, 28 (93.3 %) in cluster 2, 2 (28.6 %) in cluster 3, 15 (78.9 %) in cluster 4, 2 (40.0 %) in cluster 5.

3.2. Cluster 2: small-scale farmers, and moderate informal users of ABs, investing in phytotherapy as an alternative

Cluster 2 also includes small-scale multi-species farmers, with an average herd size of 3.3 TLU. These farmers focus mainly on chicken farming (57 %, Fig. 1A, B) and consider livestock farming as a secondary source of income (80 %, Table 1). Half of the farmers in this cluster are women, and most of them have had no education beyond primary school and do not own a fridge at home (Table 1). A majority of them (66.7 %) cite animal diseases as a difficulty for their livestock activity (Table 2). 86.7 % of farmers administer antibiotics to their animals, at an average cost of 21,272 FCFA/UBT/year, the lowest of the 5 clusters, mainly for the treatment of sick animals (67 %) (Fig. 1C, D). These antibiotics are most frequently obtained from informal sources (36.7 %) and, to a lesser extent, from veterinary pharmacists (33.3 %) (Fig. 2A). These farmers rely heavily on their personal expertise (37 %) and advice from vets (30 %) (Fig. 2B). This cluster is characterized by a widespread knowledge and use of alternatives to antibiotics (93.3 %), mainly phytotherapy, which is used by 87 % of farmers (Fig. 3A and B). The main reasons for adopting these alternatives include their efficacy (47 %) and reducing the cost of antibiotics (23 %, Fig. 3C). A significant proportion of farmers (33.3 %) saw no barrier to the adoption of alternatives, while 30 % had reservations about their efficacy (Fig. 3D).

3.3. Cluster 3: dog farmers, intensive and formal users of ABs

This cluster includes farmers specialized in dog breeding practiced by 6 out of 7 of them. Dog breeding was intended at producing dogs used for companionship. These farms are small-scale, with an average size of 0.8 TLU (Fig. 1A, B). Livestock farming is mainly a secondary source of income (57 %, Table 1), but a higher proportion of farmers consider it to be their main source of income compared to clusters 1 and 2. The farmers in this cluster are all men. Most of them have a fridge at home (71 %) and attended secondary school (85.7 %). All farmers administer ABs to their animals, supplied exclusively by veterinary pharmacists (Fig. 2A). These farmers' expenditures on ABs are high, averaging 231,590 FCFA/UBT/year (Fig. 1D). Antibiotics are mainly used for prophylaxis (57 %) and disease treatment (86 %) (Fig. 1C). Compared to farmers in the other clusters, these farmers mentioned a variety of ways of managing health problems including, in addition to medical treatment (71 %), vitamin administration (71 %), deworming (57 %), hygiene (57 %) and calling on animal health professionals for disease treatment (57 %) (Table 2). They obtain advice mainly from vets and LVAs (57 % each, Fig. 2B). They have limited knowledge and uptake of alternatives to antibiotics (29 % each). 28.6 % are aware of and use vaccination and phytotherapy (Fig. 3A and B), motivated by their effectiveness (14 %, Fig. 3C). The main obstacle to the adoption of alternatives mentioned by these farmers was their limited effectiveness (14 %, Fig. 3D).

3.4. Cluster 4: large-scale ruminant farmers, moderate and formal users of ABs

This cluster includes large-scale ruminant farmers, with an average size of 24 TLU (Fig. 1A, B), the largest among the clusters. For these farmers, who are almost all men (94.7 %, Table 1), livestock production is most often the main source of income (74 %, Table 1). Most of them were not educated beyond primary school and did not own a fridge at home (Table 1). A majority of them (52.6 %) considered animal diseases to be a problem for their livestock (Table 2). All of them administer ABs, mainly supplied by veterinary pharmacists (95 %) (Fig. 2A) and mainly intended for disease treatment (84 %, Fig. 1C). These farmers spend moderately on ABs, with an average of 81,591 FCFA/TLU/year (Fig. 1D). They turn to vets for advice on the use of ABs (53 %, Fig. 2B). They are characterized by a high level of knowledge (79 %) and practice (63 %) of alternatives to antibiotics. These alternatives mainly include phytotherapy and hygiene, known by 58 % and 37 % of farmers respectively, and practiced by 47 % and 32 % respectively (Fig. 3A and B). The main reasons for adopting these alternatives are the high cost of antibiotics (26 %) and the effectiveness of the alternatives (21 %, Fig. 3C). The main obstacle to the use of alternatives mentioned by these farmers is their limited effectiveness (47 %, Fig. 3D).

3.5. Cluster 5: commercial poultry farmers, intensive and formal users of ABs

This cluster includes large-scale poultry producers, with an average size of 13 TLU (Fig. 1A, B). Like the dog farmers, the majority (80 %) have a fridge at home and all of them attended secondary school. Unlike the poultry farmers in the other clusters, these farmers use exotic breeds of poultry rather than local breeds. Another specific characteristic of these farmers is their use of at least two barns and the practice of sanitary emptying between two flocks (Table 2). For these farmers, who are all men, livestock farming is the main source of income (Table 1). All of them administer ABs, supplied exclusively by veterinary pharmacists (Fig. 2A). These farmers present a high level of expenditure in ABs, averaging 265,981 FCFA/TLU/year, but with great disparities across farmers (Fig. 1D). ABs are used for a variety of reasons, including disease treatment (40 %), metaphylaxis (40 %) and prophylaxis (60 %) (Fig. 1C). For health management, the majority of farmers prefer medicinal treatments (80 %, Table 2). Farmers almost exclusively consult vets for advice (80 %, Fig. 2B). These farmers have little knowledge or practice of alternatives to antibiotics (40 % for each). 40 % of them mentioned being aware and practicing phytotherapy (Fig. 3A and B). The obstacle to the adoption of alternatives most often mentioned by these farmers was their limited effectiveness (40 %, Fig. 3D).

3.6. Characteristics of distributors

The interviewed distributors logically fall into two distinct groups with very clear differences (Appendix XI): 12 formal distributors (including 8 veterinary pharmacists and 4 LVAs) and 32 informal distributors (11 market retailers, 20 local retailers, and one informal vendor with an atypical profile). Local retailers are generally located in the same neighborhood as the farmers but outside markets. The vast majority of formal distributors are men (11 out of 12), while the vast majority of informal distributors are women (29 out of 32). Among the formal distributors, 67 % have no other source of income apart from medicine sales. 50 % are mainly supplied by one of the veterinary pharmacies and 42 % are supplied by legal imports, the veterinary pharmacies themselves being entirely supplied by legal imports from foreign firms. They mentioned distributing medicines to livestock farmers (100 %), vets (58 %), LVAs (58 %) and pet owners (42 %). Conversely, 72 % of informal distributors have another source of income. 72 % of them are mainly supplied by informal wholesalers in the markets. Unfortunately, all contacted wholesalers declined to be interviewed. According to the information gathered from marker retailers, the wholesalers import the medicines from neighboring countries of Togo through illegal cross-border trade. 88 % of informal distributors do not know the profession of a large fraction of their customers or the purpose of the medicines sold (used on humans or animals), but 25 % are aware that their customers include livestock farmers. In addition to ABs, all the formal distributors sell feed supplements and dewormers, and a majority sell vaccine (58 %). In contrast, informal distributors mainly sell antibiotics for human use (94 %). In terms of knowledge of alternatives to antibiotics, formal distributors are aware of a variety of alternatives: 50 % of them cite vaccination as an alternative, followed by phytotherapy and disinfection (8 % each). In contrast, informal distributors are only aware of phytotherapy as an alternative to ABs, known by 63 % of them. All formal distributors are aware of problems associated with the use of ABs, while 81 % of informal distributors are not aware of any problem. Seven formal distributors and 3 informal distributors mentioned the lack of efficacy of the drugs as a problem associated with AB use. Six formal distributors and one informal distributor mentioned the misuse of ABs, including not respecting treatment durations and dosages. Four formal distributors and 1 informal distributor mentioned animal mortality or deteriorating animal health due to overdosing. All formal distributors identified obstacles to the use of alternatives and the majority (58 %) identified the lack of information communicated on these alternatives to farmers as a major barrier to their use. 50 % of the formal distributors mentioned other obstacles such as (i) the lack of financial resources, which was mentioned several times, (ii) the reluctance of farmers to use alternatives and to follow veterinary recommendations, and (iii) the difficulty of accessing alternative products. 80 % of informal distributors did not perceive any obstacle to the use of alternatives and 15 % mentioned the lack of information as an obstacle.

3.7. AB distribution channel

We chose to analyze the origin of the ABs used by farmers in clusters 1 and 2 (the small-scale moderate and informal AB users, hereafter referred to as group 1) separately from clusters 3, 4, and 5 (commercial formal AB users, hereafter referred to as group 2). Indeed, these two groups have distinct purchasing behaviors: while farmers in Group 1 most often obtain their ABs from informal actors (market retailers and local retailers), almost all farmers in Group 2 obtain their ABs from veterinary pharmacies. According to the results, 79 % of ABs purchased by farmers of group 1 came from legal imports supplying veterinary pharmacies, 2 % from human pharmacies, and 19 % from market wholesalers, who are likely supplied by illegal cross-border trade (Fig. 4A). Regardless of their origin (legal imports, illegal imports, or human pharmacies), 4 % of ABs transited through animal health professionals (vets, technicians, or LVAs), 79 % through veterinary pharmacists, 23 % through market retailers and less than 1 % were distributed by local retailers. Of all the ABs purchased by farmers in group 2, 99 % came from legal imports, while less than 1 % came from human pharmacies and market wholesalers (Fig. 4B). 99 % of ABs were purchased from veterinary pharmacies and less than 1 % from animal health experts and market retailers.

Fig. 4.

Antibiotic distribution circuit supplying the farmers in the clusters identified on the basis of survey data from the urban area of Lomé (February 2023–March 2024). Diagrams A and B describe the supply of antibiotics to farmers in clusters 1 and 2 and to farmers in clusters 3, 4, and 5 respectively. Categories of distributors involved in the antibiotic supply are displayed with boxes with colors indicating the actors belonging to the formal veterinary (green), human health (yellow), or informal (red) sector. The width of the arrows indicates the relative quantitative importance of the flows of antibiotics between the categories of actors and the fractions of antibiotics distributed to farmers transiting through each category are indicated in percentage on the right of the category concerned. *Animal health expert: includes livestock veterinary auxiliaries (LVAs), veterinarians, and animal health technicians. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

4. Discussion

The results of our study highlight the diversity of AB uses and modalities of AB supply to livestock farms in the urban area of Lomé, and their linkages with the socio-economic characteristics of actors involved in livestock farming and in the ABs' distribution.

The highest expenditure on ABs in relation to animal biomass was observed among dog farmers and commercial poultry producers. This high consumption is explained by the widespread use of ABs for prophylactic or metaphylactic purposes in these farms, whereas small-scale mixed and large-scale ruminant farmers essentially restrict ABs to therapeutic use. These two categories of intensive AB users are distinguished by their higher socio-economic status, indicated by their high level of education and the presence of a refrigerator at home. Poultry farmers have intensive farming characteristics, with large, single-species farms being their main source of income, at least two buildings per farm, and the use of exotic breeds. Dog farms, on the other hand, are characterized by their very small size, which is probably compensated by the high commercial value of the reared animals. It is generally accepted that the larger the herd, the more inclined farmers are to implement preventive practices [28]. While several commercial poultry producers practice sanitary emptying, only a few opt for vaccination, and none have indicated that they disinfect their premises and equipment. They resort to the intensive use of ABs to secure their production and income, despite the financial cost involved. The same observation was made on farms with similar characteristics in Senegal and India well [29,30]. Rware et al. (2024) and Boamah and Agyare (2016) also showed that the intensity of the use of ABs increases with the size of poultry farms in Kenya and Ghana respectively [12,19]. It should be noted that a majority of these intensive users of ABs declare seeking advice from animal health professionals, including veterinarians, from whom they source ABs, and that dog farmers seek the assistance of animal health services to help them treat sick animals. The persistence of excessive use of ABs, despite the expertise provided by these qualified players, calls into question the role they may play in accompanying farmers towards a more rational use of ABs. Animal health professionals may, in some cases, encourage the use of ABs because of (i) a preference for ABs stemming from their personal experience, (ii) their economic motivation to sell off their stocks of ABs, (iii) the desire to minimize the risk of health complications and dissatisfaction of their customers, and (iv) pressure from some animal owners to prescribe ABs [31].

Small-scale mixed farmers have a more moderate use of ABs, most of which are administered to treat the sick rather than for prevention. This can be explained by (1) the secondary role of livestock farming in their income, which makes them less economically vulnerable to production losses due to disease, and (2) the widespread use of phytotherapy. This reflects a desire to minimize the cost of medicinal treatments, given the low cost and local availability of medicinal plants, the effectiveness of which seems to satisfy the majority of these farmers. The large number of women in this category can be considered an additional explanation. Since, according to previous studies, women have more limited control over households' financial resources and less access to veterinary services, they are more likely to use local resources and low cost products [32]. Compared with the intensive AB users, a high proportion of these farmers said they relied on advice from other farmers or their own experience to diagnose diseases and decide on treatments, rather than on advice from animal health professionals. As farmers' knowledge is largely based on empirical observations, this can lead to errors, particularly as a result of confusion between symptoms caused by different pathogens [33]. Soro et al. have identified the risks associated with the inappropriate use of ABs and non-compliance with dosages on farms in Burkina Faso [33].

On large-scale and canine farms, almost the entirety of administered ABs are veterinary medicines supplied by legal imports and are obtained from veterinary pharmacists, who are the central players in the formal AB distribution network. This purchasing behavior was also reported by Soro et al. (2024) in Burkina Faso and by Morang'a et al. (2024) in Kenya [33,34]. The majority of small-scale farmers obtain their ABs through informal channels, in which ABs for human use circulate, and whose central players are wholesale and retail sellers in urban markets. ABs sold in informal channels apparently come from illegal cross-border trade. However, our study estimates that the majority of ABs used in small-scale farms still come from legal imports. This apparent contradiction is due to (1) the larger quantity of ABs used by farmers who buy their ABs from formal distributors and (2) a minor fraction of ABs sold by market retailers coming from veterinary pharmacies. This estimate may be biased, however, because expenditure on ABs was used as an indicator of the quantity of ABs purchased by farms, without taking into account the price differences that may exist between ABs sold through formal and informal channels. Large farms tend to favor formal veterinary pharmacies probably because they offer a dependable source of trustworthy products and guidance on managing animal health. Conversely, small-scale farmers opt for informal distributors, probably because of their ease of access and their comparatively lower sale prices. Medicines purchased in the informal sector can be obtained at any time, without consulting a health specialist, whose services may be financially inaccessible to these farmers with limited resources. Another potential explanation is the lack of knowledge about the risks associated with antibiotic misuse and the purchase of counterfeit or under-dosed products [35]. We can also assume the existence of sociological barriers to accessing more qualified actors, in relation to the farmers' low socio-economic status or their gender. The purchase of ABs in the informal sector increases the risk of unsuitable AB use or the use of critical ABs reserved for human medicine. While formal distributors are aware of issues associated with the use of ABs, which corroborates the results of other studies conducted with these same types of actors [36], informal distributors, who are not health professionals, displayed more limited knowledge of the topic.

While the groups of intensive users and formal distributors of ABs are almost all men, women make up a large proportion of small-scale mixed livestock farmers and the vast majority of informal distributors of ABs. This unequal gender distribution across categories of actors is to be linked with the broader socioeconomic context of sub-Saharan African countries, where women have limited access to education [37] and, consequently, to funding and entrepreneurship, compared to men [38]. Additionally, studies carried out in West Africa showed that women play an important role in the informal cross-border trade of goods [39,40], due to their marginalization and exclusion from more formal commercial exchanges [41]. Although women's roles in livestock production are influenced by cultural norms, which tend to limit their involvement to specific tasks such as cleaning pens and feeding livestock [42], some studies show that improved access to education translates into an increased involvement of women in livestock management and marketing [38,42].

Numerous studies have highlighted the benefits of promoting alternatives to ABs for controlling bacterial diseases while, at the same time, limiting AMU on farms [43]. However, their effectiveness depends to a large extent on how well they are understood and used by farmers, which requires clear and appropriate communication. Several potential alternatives to ABs were mentioned by participants in the survey, with varying degrees of awareness and adoption, including, in decreasing order of frequency, phytotherapy, hygiene and biosecurity measures, vaccination, and appropriate feeding. However, several obstacles to the adoption of alternatives were identified, including (1) farmers' lack of knowledge about these alternatives, observed by the formal distributors interviewed and corroborated by the farmers' interview results, and (2) farmers' uneven perception of their effectiveness. While a large proportion of small-scale mixed farmers use phytotherapy and do not perceive any obstacles to its use, this is not the case for dog farmers and large-scale poultry and ruminant farmers, for whom the limited effectiveness of these treatments appears to be a major obstacle to their use. Studies evaluating the quality of plant-based medicines pointed to the difficulty of guaranteeing a stable and reproducible composition of these medicines and that their chemical composition can vary depending on their origin, growing conditions, and processing methods, making it difficult to ensure their quality and market them [44]. Additionally, the effectiveness of phytotherapy depends on the animal species and health status [45]. Only a minority of farmers use vaccination to prevent infectious diseases, regardless of the category. Dog farmers more frequently cite vaccination as an alternative, but they probably use it for preventing specific viral diseases such as rabies and parvovirus. Although disease prevention by vaccination is an effective alternative to ABs, vaccines have the disadvantage of being specific to certain pathogens for which vaccines are developed and marketed. Strengthening biosecurity measures theoretically enables an effective prevention of bacterial infections, thereby reducing the need for AB administration [46]. However, many biosecurity measures can only be implemented in a confined environment and require substantial investment or major structural changes in farming systems. They therefore seem more accessible to commercial dog and poultry farmers, who are the largest consumers of ABs, and to a lesser extent to large-scale ruminant farmers, who have more extensive farming practices. However, commercial farmers, like all the other categories, pointed to financial constraints as a major difficulty of their activity, which calls into question their ability to invest financially in biosecurity. Optimizing animal feed is another way of strengthening the health of animals while reducing the need for ABs [47,48], but it was hardly mentioned by farmers.

Our study has some limitations. Although the questions were asked in the local language or in French in order to maximize their comprehensibility for the participants, translation and interpretation biases cannot be ruled out. In addition, respondents sometimes had to rely on their memory of events dating back a year or more, which could introduce memory bias. Given the informal nature of the activity of many actors of domestic animal farming systems and AB distribution, snowball sampling is the only reliable way of identifying the actors of the AB distribution network in a comprehensive way. However it may have led to an over-representation of people with strong social ties who shared similar characteristics, which may limit the generalizability of the results to the entire population [49]. The interviewed sample of farmers cannot be considered representative of the population of farmers of the Maritime region, since living areas of MRB carriers were targeted on purpose. Therefore, our quantitative measures cannot be generalized, as such, to the entire population.

5. Conclusion

Our results demonstrate the need to adapt interventions aimed at preventing the misuse of ABs to issues specific to each category of farmers and their corresponding AB distribution channels, which implies the adoption of a gender-based approach. Farmers who use ABs excessively are mainly men of relatively high socioeconomic status who practice a commercial farming of dogs and poultry and are supplied in ABs by the formal distribution circuit. These farmers are particularly sensitive to the effectiveness of animal disease management methods but are probably the most capable of investing financially in alternatives to ABs, such as biosecurity and vaccination, provided they receive adequate financial support, which could take the form of a targeted credit program. These investments need to be informed and supported by pharmacists, vets, and technicians, who are the preferred partners of these farmers when it comes to managing animal health, and who themselves need to receive specific training on this issue. Farmers who make moderate or occasional use of ABs, a large proportion of whom are women, obtain their ABs in the formal and informal circuits, the latter being mainly supplied with ABs for human use. These farmers, who often keep livestock extensively and as a secondary activity, unfrequently seek support from animal health professionals. They are nonetheless familiar with and invest in alternatives to ABs, primarily herbal medicines, which are readily available and at low cost. The potential misuse associated with the purchase and use of ABs outside formal channels by these farmers could be effectively addressed by targeted communication campaigns on the risks associated with non-prescribed, counterfeit, or incorrectly dosed medicines, and the use of ABs that are critical to human health. These campaigns should target both small-scale farmers and informal distributors and take into account the gender issues and barriers to access to veterinary services, which could potentially discourage the actors of this category from implementing the promoted changes.

CRediT authorship contribution statement

Esso-tchella Madera Bodombossou: Writing – original draft, Methodology, Investigation, Formal analysis, Data curation. Komi Agbessi Adjessoklou: Investigation. Rogatien Comlan Atoun: Validation. Isidore Tchaou: Validation. André Pouwedeou Bedekelabou: Validation. Dossou Zanou: Validation. Gad Boukaya: Validation. Camille Akapko: Validation. Laurence Armand Lefevre: Validation. Mounerou Salou: Validation. Gilles Brücker: Validation. Didier Koumavi Ekouevi: Validation. Dominique Salmon-Céron: Validation. Andrea Apolloni: Validation, Methodology. Eric Cardinale: Validation, Project administration, Methodology, Conceptualization. Alexis Delabouglise: Writing – original draft, Validation, Methodology, Formal analysis.

Funding

The study was supported by the ONE HEALTH TOGO (ONHETO) which funded by the French Ministry of Solidarity and Health [grant number 22-SB4824].

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

Supplementary material 1

Supplementary material 2

Data availability

The data was collected, stored, processed, and shared among members of the research team directly involved in data analysis, in compliance with the European Union's General Data Protection Regulation (2016/679). The original database of the study will be made freely available after full anonymization (total removal of personal data of participants), before 2030. However, this database may be shared if an explicit request is made to: Eric Cardinale (eric.cardinale@anses.fr).