This study, conducted in Niamey, Niger, combined qualitative methods (observations and interviews) and quantitative methods (laboratory analyses) to assess the microbiological quality of Kilishi (the finished product). Sample collection took place from June to September 2024, and analyses ran from October 2024 to February 2025.

Data were collected through semi-structured interviews and direct observations with 60 actors in the Kilishi value chain (producers and sellers). Interviews focused on production stages (meat selection, drying/cooking techniques, preservation, and packaging) (Figure 1). Special attention was paid to hygiene conditions and the production environment (exposure to dust, insects, and stray animals), with visual documentation produced.

Figure 1. Kilishi production diagram.

Sixty (60) Kilishi samples were collected from vendors, following their standard cutting and packaging procedures. Samples were placed in sterile freezer bags, labeled, and transported promptly in a cooler with cold packs to the LaBESTA Laboratory in Ouagadougou, Burkina Faso, for microbiological analysis.

Microorganisms were detected using the standard solid culture technique.

2.4.1. Solution and Dilution Preparation

A stock solution (pre-enrichment) was prepared by placing 25 g of Kilishi in 225 ml of Buffered Peptone Water (BPW) (Liofilchem), incubated at 37˚C for 24 h. Decimal dilutions were performed up to 10⁻³ for Salmonella and 10⁻⁶ for Escherichiacoli (E. coli).

2.4.2. Detection and Identification of Salmonella spp.

Detection followed standard 6579-1: Selective Enrichment: 0.1 ml of pre-enriched broth was plated on Rappaport Vassiliadis (RV) broth at 42˚C, and 1 ml was added to Muller-Kauffman Tetrathionate (MKTT) broth at 37˚C, both incubated 18 to 24 h.

Isolation: Enriched broth was streaked onto selective SS and XLD agar (Liofilchem).

Identification: Suspect strains were identified using the API 20E biochemical profile (BioMérieux).

2.4.3. Serotyping of Salmonella spp.

Serotyping, following biochemical identification, was performed by direct slide agglutination with specific antisera.

2.4.4. Enumeration of Thermotolerant Coliforms and E. Coli Identification

Coliforms were enumerated by surface plating on Violet Red Bile Lactose (VRBL) medium, incubated at 44˚C for 24 h (Standard V08-060, AFNOR, 2009) [7]. Colonies between 15 and 150 were counted to calculate the bacterial load (CFU/gram) using the standard formula. E. coli suspects were confirmed using the API 20 E biochemical gallery (BioMérieux).

N = ∑ C ( n 1 + 0 , 1 n 2 ) * d * V

N = Number of colony-forming units (CFU)/gram of food;

V = Volume of solution deposited;

ΣC = Total number of colonies counted in boxes with colonies between 15 and 150;

n₁ = Number of boxes counted from the first dilution;

n₂ = Number of boxes counted from the second dilution;

d = Dilution factor from which the first counts were made.

2.4.5. Multiplex Polymerase Chain Reaction (16-Plex PCR)

This study employed a 16-plex PCR for the simultaneous detection of 16 virulence genes belonging to the five main E. coli pathovars (Enterohemoragic Escherichiacoli (EHEC), Enteropathogenic Escherichiacoli (EPEC), Enteroaggregative Escherichiacoli (EAEC), Enteroinvasive Escherichiacoli (EIEC), Enterotoxinogenic Escherichiacoli (ETEC) [8].

DNA Extraction: Bacterial DNA was extracted from pure E. coli strains using the heating method (boiling for 10 min, followed by centrifugation at 11,337 rpm for 10 min). The supernatant was stored for PCR.

Amplification: The 20 µl reaction included a premix (Taq polymerase, dNTPs, buffers, MgCl₂) and specific mixtures of 16 primers (Mix 1 and Mix 2). The thermocycler program involved 35 cycles with a hybridization step at 62.5˚C for 60 seconds.

Electrophoresis: PCR products were separated on a 2% agarose gel, stained with ethidium bromide, and visualized under a UV lamp. A 100 bp molecular weight marker was used.

Interpretation: Pathovar determination was based on the presence of specific genes (elt for ECET; stx1/stx2 for ECST; uidA as the general E. coli marker) (Table 1). Reference E. coli strains from THL (Helsinki, Finland) served as controls.

Table 1. Primer sequences for multiplex PCR.

STEC: Shigalike Toxin E. coli; ETEC: Enterotoxinogenic E. coli; EIEC: Enteroinvasive E. coli; EAEC: Enteroaggregative E. coli; Ref.: Reference; T: Expected amplicon size in base pairs (bp); [C]: Concentration; 1: [9]; 2: [8].

2.5.1. Qualitative Analysis

Data from interviews and observations were transcribed and subjected to qualitative thematic analysis to identify key manufacturing stages and potential health risks.

2.5.2. Quantitative and Statistical Analysis

Quantitative microbiological data underwent descriptive statistical analysis to determine the prevalence of the different microorganisms (expressed as frequencies and percentages). Software: Excel and Epi-Info version 3.5.1 were used for prevalence calculations. Statistical Significance: Med Calc 11.0.1.0 software was used to determine p-values. Differences were considered statistically significant at a p-value less than or equal to 0.05.

Artisanal Kilishi in Niamey is the result of a traditional process, involving specific steps to guarantee its texture, taste, and preservation.

3.1.1. Meat Selection and Preparation

Kilishi is mainly prepared from beef, with the hind leg being preferred for its yield. The meat is first scraped clean of bone fragments, then meticulously flattened into thin strips to ensure optimal and uniform drying.

3.1.2. Initial Drying

Meat slices are spread out for intense sun drying, usually in the open air. Observations have shown that this phase lasts about 3 hours per side, after which the meat is turned over. The drying process lasts approximately 24 hours. During this stage, the meat is exposed to environmental elements, which is a critical point for product hygiene and safety.

3.1.3. Preparation, Seasoning, and Final Cooking

After initial drying, the meat is coated with peanut oil and seasoned. Salt and peanut oil are essential ingredients found in all varieties. Three main varieties of Kilishi exist. Red Kilishi: Distinguished by the addition of red chili pepper (Tattassé), White Kilishi: Includes classic condiments (groundnuts, garlic, ginger, chilies) but without red pepper, and Unseasoned Kilishi: Contains only peanut oil and salt.

The seasoning mixture is evenly spread. The meat is then grilled over a wood fire at a low temperature (about 15 to 20 minutes per side. This final cooking provides the Kilishi its unique flavor and appearance while potentially reducing the existing microbial load.

3.2.1. Shelf Life and Storage

Kilishi is known for its long shelf life, ranging from three months to one year at ambient temperature without refrigeration, attributed to low water activity resulting from intensive drying and cooking.

3.2.2. Packaging

The product is typically wrapped in paper (including paper from empty cement bags) or aluminum foil. Plastic bags are avoided. The use of recycled paper raises hygiene concerns.

3.2.3. Display Conditions and Sales Points

Sales occur in various locations (markets, unofficial stalls) with varied exposure. Observations highlighted: Frequent open-air display, sometimes only covered by a piece of cloth. Critical Factors: Direct exposure to dust, high ambient temperatures, constant presence of insects (flies), and, in some cases, proximity to stray animals. Handling: Frequent direct hand handling by vendors without gloves poses a potential source of cross-contamination post-cooking.

The collected Kilishi samples showed contamination primarily with Salmonella and E. coli, with frequencies of 4 (6.7%) and 25 (41.7%), respectively (Figure 2). The microbial loads of these strains were quantified and are presented in Table 2. Regarding faecal coliform counts, the bacterial load ranged from 2 × 10² CFU/g to 1.20 × 10⁶ CFU/g, with an average of 3.40 × 10⁴ CFU/g.

Figure 2. Prevalence of bacteria in kilishi samples

Table 2. Microbial load of thermotolerant coliforms and Salmonella in samples.

N = Number of sample, p = 0.40 > 0.05. The ‘a’ in bold indicates that there is no significant difference between the results for the different sectors.

3.3.1. Prevalence of Salmonella Serovars

The results show a variable prevalence of Salmonella in the different kilishi meats collected from the sales sites in the city of Niamey. It appears that 6.7% (4/60) of the samples are contaminated with Salmonella spp. The study also shows that 1.67% are infected with Salmonellatyphi (Table 3).

Table 3. Prevalence of Salmonella serovars.

N = Total number of Salmonella isolated; % = Percentage.

3.3.2. Prevalence of E. coli Pathovars

Table 4. Prevalence of E. coli pathovars.

N = Total number of Salmonella isolated; ECST = Shiga-like toxine Escherichiacoli; ECEA = Enteroaggregative Escherichiacoli; ECEI = Enteroinvasive Escherichiacoli; ECET = Enterotoxinogenic Escherichiacoli.

Figure 3. Agarose gel image showing multiplex PCR amplicons.

Molecular analyses showed that 41.7% of Kilishi intended for human consumption contained E. coli strains associated with diarrhoea. Five E. coli pathovars were identified. The study showed a prevalence of 6.67% ECST, 5% ECST-ECET, 0% ECEA, 0% ECEI, and 1% ECET. The highest prevalence rates were observed for ECST and ECST-ECET (Table 4). The uidA gene confirms that the E. coli strain was identified in all strains tested. These strains, which only carry the uidA gene, are classified as non-diarrhoeic E. coli (Figure 3).