Resistant Bacteria Turn Fresh Vegetables Into a One Health Surveillance Gap in Burkina Faso
Fresh vegetables sold and eaten raw in Ouagadougou carried a substantial bacterial burden, and the pattern was not a seasonal fluke. In 300 samples drawn from farms, markets, restaurants, street-food settings, and supermarkets, Escherichia coli appeared in 63.0% of specimens and Salmonella spp. in 10.0%. Those organisms matter far beyond food hygiene checklists: they are common enteric pathogens, and when they travel with antimicrobial resistance determinants, they become a harder clinical problem once infection occurs.
The study places fresh produce squarely inside the antimicrobial resistance conversation that has often centered on hospitals, livestock, and wastewater. Raw vegetables do not receive a kill step before consumption, so contamination acquired in the field, during transport, at market, or in food-service preparation can survive all the way to the plate. In settings where cold storage is limited and washing practices are inconsistent, the route from soil and water to human ingestion is short.
Contamination Was Common Across the Supply Chain
Sampling covered the supply chain in a way that reflected how vegetables move through the city. Farm sites, open-air markets, restaurants and street-food establishments, and supermarkets were all represented, with lettuce, tomato, cucumber, onion, and mixed salads included among the products. The distribution revealed that fresh produce contamination is not confined to one node in the chain; it can emerge before harvest and persist through retail and service environments.
Seasonal comparison did not reveal a meaningful shift in prevalence. E. coli rates were 61.3% in the dry season and 64.7% in the rainy season, while Salmonella rates were 8.0% and 12.0%, respectively. The absence of a clear seasonal separation suggests that the contamination pressure is structural rather than episodic, likely sustained by handling practices, ambient storage, and repeated exposure to contaminated surfaces or water.
Resistance patterns followed the same general logic. Per-agent resistance rates were broadly similar between dry and rainy seasons, indicating that weather alone does not explain the microbiology seen in the sampled vegetables. The problem is embedded in the supply environment.
Salmonella Showed Heavier Resistance Pressure Than E. coli
Across individual drugs, Salmonella isolates were more resistant than E. coli to several agents, especially tetracycline and trimethoprim-sulfamethoxazole. Tetracycline resistance reached 25.0% in Salmonella versus 15.3% in E. coli, and trimethoprim-sulfamethoxazole resistance was 21.7% versus 12.8%. Those differences are clinically relevant because they point to a reservoir of organisms that have already survived exposure to commonly used antimicrobial classes.
Multidrug resistance deepened the concern. Among resistant Salmonella isolates, 60% met the definition of multidrug resistance, meaning non-susceptibility to at least three antimicrobial classes. In practice, that pattern raises the probability that a foodborne infection could be less responsive to first-line empiric therapy, especially where laboratory susceptibility testing is not immediately available.
The resistance profile also fits with the ecological reality of food systems in West Africa. Antibiotic selection pressure in human medicine, veterinary care, and agriculture can converge in the same environmental compartments. Fresh produce becomes a carrier not only of organisms, but of the resistance traits those organisms have accumulated.
blaCTX-M Appeared in Vegetable-Associated Isolates for the First Time in This Surveillance Context
Eight resistant isolates were selected for multiplex PCR screening, four E. coli and four Salmonella, chosen to represent the range of resistance phenotypes observed. In that subset, blaTEM appeared in all eight isolates, blaOXA in 50.0%, and blaCTX-M in 50.0%. Three isolates carried all three genes.
The presence of blaCTX-M is the most important genetic finding in the report. CTX-M-type extended-spectrum beta-lactamases can compromise third-generation cephalosporins, which remain important treatment options for severe enteric infections. Detecting that gene in bacteria associated with fresh vegetables expands the surveillance frame in Burkina Faso beyond human specimens, livestock, and water. It adds the agricultural-food interface to the map of circulating resistance genes.
A signal like this does not prove transmission into patients by itself, but it shows the biological route is available. Resistance genes move through mobile genetic elements, and fresh vegetables can serve as a junction where environmental, animal, and human bacterial populations overlap.
Handling Practices Created an Easy Path for Cross-Contamination
Food handling conditions were poor in ways that directly support bacterial persistence. Nearly all vendors, 98%, stored vegetables at ambient temperature. Disinfection was absent in 85% of settings. Cross-contamination risk with raw meat was present in 80% of settings.
Those figures are not abstract process gaps; they are the operational conditions that let enteric bacteria survive on foods eaten raw. Ambient storage reduces the chance of limiting bacterial growth. Lack of disinfection leaves surface contamination intact. Contact with raw meat introduces another route for fecal organisms and resistant bacteria to move into salad ingredients that may never be cooked.
A screening-level exposure model estimated a 2.0% risk for antibiotic-resistant Salmonella per raw salad serving. The manuscript clearly frames this as a screening indicator rather than a confirmed infection probability, so it should be read as a signal for prioritization, not as a direct incidence estimate. Even so, the model reinforces the practical point: one serving can carry a non-trivial microbial exposure burden when sanitation is weak.
HACCP Control Points Offer a Starting Framework, Not a Finished Solution
The investigators proposed a HACCP-based control framework with five critical control points tailored to the local supply chain. That proposal is useful because the failure points are now visible: contamination at source, handling during transport and sale, preparation in food-service settings, and the absence of effective decontamination before consumption.
The framework still needs field validation. The study used purposive sampling rather than probability sampling, and the authors are careful to frame the results as descriptive of the sampled venues rather than fully generalizable to the entire city or country. The point is narrower but important: fresh produce is already part of the AMR burden in Burkina Faso, and surveillance systems that ignore vegetables will miss a meaningful environmental reservoir.
The broader implication reaches beyond one city. Fresh vegetables are eaten widely, often raw, and they sit at the intersection of agriculture, food service, and household consumption. In West Africa, where AMR surveillance has largely focused on clinical specimens and a few food categories, vegetables belong in the monitoring strategy now, not later.