An international multicenter study of ready-to-eat foods, sandwiches, and ice creams or sorbets sold in the streets and their vendors was carried out to assess the microbiological quality of these foods and to identify characteristics of the vendors possibly associated with pathogens. Thirteen towns in Africa, America, Asia, and Oceania were involved in the study. A single protocol was used in all 13 centers: representative sampling was by random selection of vendors and a sample of foods bought from each of these vendors at a time and date selected at random. Microbiological analyses were carried out using standardized Association Française de Normalisation methods, and the use of a standardized questionnaire to collect data concerning the characteristics of the vendors. Fifteen surveys were carried out, with 3,003 food samples from 1,268 vendors. The proportion of unsatisfactory food samples was between 12.7 and 82.9% for ice creams and sorbets and between 11.3 and 92% for sandwiches. For ice creams and sorbets, the sale of a large number of units (>80 per day) increased the risk of unsatisfactory food by a factor of 2.8 (95% confidence interval [CI]: 1.5 to 5.1), lack of training in food hygiene by 6.6 (95% CI: 1.1 to 50). and by a factor of 2.8 (95% CI: 1.4 to 5.4) for mobile vendors. These risk factors were not identified for sandwiches, this difference may be due to the presence of a cooking step in their preparation. These results show that the poor microbiological quality of these street foods constitutes a potential hazard to public health, that the extent of this hazard varies between the cities studied, and that vendors' health education in food safety is a crucial factor in the prevention of foodborne infections.

A total of 127 strains of Vibrio cholerae (117 V. cholerae O1 and 10 nonagglutinating strains) isolated from a recent cholera outbreak in Senegal and four strains isolated in Guinea-Bissau (during the survey of a cholera epidemic that occurred 10 months before the Senegalese one) were analyzed. Strains were characterized by conventional methods (biochemical and serologic identification, susceptibility to antimicrobial agents), polymerase chain reaction for genes encoding cholera toxin (CtxA), zonula occludens toxin (Zot), and accessory cholera enterotoxin (Ace), and by ribotyping. Conventional methods showed that all strains of V. cholerae O1 belonged to serotype Ogawa, biotype El Tor and were resistant to the vibriostatic agent O129 (2,4-diamino 6,7-diisopropylpteridine phosphate), cotrimoxazole, and chloramphenicol; all strains were sensitive to tetracycline, a drug that has been extensively used in cholera therapy. Most of these V. cholerae O1 (112 strains from Senegal and four strains from Guinea-Bissau) had an intact core region (virulence cassette) and amplified a 564-basepair (bp) fragment of ctxA, a 1083-bp fragment of zot, and a 314-bp fragment of ace. Ribotyping of V. cholerae O1 strains after Bgl I restriction of total DNA revealed that ribotype B5a, which is the predominant ribotype of this seventh pandemic of cholera, was not isolated. Instead, a new ribotype was identified and designated B27 in our data bank. Since O1 isolates from Guinea-Bissau and Senegal have the same biotype, serotype, and ribotype and as the Guinea-Bissau outbreak that preceded the one in Senegal, this emerging ribotype probably came from Guinea-Bissau. Nonagglutinating strains exhibited no resistance to the O129 agent and to the tested antibiotics, they were all negative for virulence cassette, except for one strain with the ctxA and zot genes isolated from a patient with diarrhea, and there was a great variability of ribotypes among these strains. There was no difference between environmental O1 strains isolated from water and strains isolated from patients with cholera, suggesting that fecally contaminated water is an important reservoir for infection.

Because the use of live attenuated mutants of Shigella spp. represents a promising approach to protection against bacillary dysentery (M. E. Etherridge, A. T. M. Shamsul Hoque, and D. A. Sack, Lab. Anim. Sci. 46: 61-66, 1996), it becomes essential to rationalize this approach in animal models in order to optimize attenuation of virulence in the vaccine candidates, as well as their route and mode of administration, and to define the correlates of protection. In this study, we have compared three strains of Shigella flexneri 5-the wild-type M90T, an aroC mutant, and a double purE aroC mutant-for their pathogenicity, immunogenicity, and protective capacity. Protection against keratoconjunctivitis, induced by wild-type M90T, was used as the protection read out in guinea pigs that were inoculated either intranasally or intragastrically. Following intranasal immunization, the aroC mutant elicited weak nasal tissue destruction compared to M90T and achieved protection correlated with high levels of local anti-lipopolysaccharide immunoglobulin A (IgA), whereas the purE aroC double mutant, which also elicited weak tissue destruction, was not protective and elicited a low IgA response. Conversely, following intragastric immunization, only the M90T purE aroC double mutant elicited protection compared to both the aroC mutant and the wild-type strain. This mutant caused mild inflammatory destruction, particularly at the level of Peyer's patches, but it persisted much longer within the tissues. This could represent an essential parameter of the protective response that, in this case, did not clearly correlate with high anti-lipopolysaccharide IgA titers.

No abstract

No abstract