Dietary habits and gastrointestinal colonization by antibiotic resistant microorganisms

Sir, The human gastrointestinal flora is affected by probiotics such as lactobacilli, the use of antibiotics and, last but not least, dietary habits. It appears obvious that intestinal metabolism and mucosal immunity change with the type of food intake, which is also correlated with changes in the ratio in which different microorganisms occur. Antibiotics, like food, greatly affect the composition of the gastrointestinal flora and may select for resistant strains. This happens not only for patients, but may also occur when antibiotics are used as growth promoters in modern foodanimal production. Meat products have been shown to be colonized with resistant microorganisms. This raises the questions of whether meat serves as a vector during zoonotic transmission, and whether antibiotic-resistant microorganisms present in food may persistently colonize the human gastrointestinal tract. Such questions could be answered by designing studies that compare the faecal flora of vegetarians and that of meat-eating controls. We performed a large study to determine the presence of vancomycin-resistant microorganisms in the faecal flora of vegetarians in 1997. We obtained rectal swabs from 318 vegetarians (mean age 55 18 years, mean number of meat free years 30 2) and 276 control individuals (mean age 53 17 years). The swabs were inoculated in Enterococcosel medium (BBL, Cockeysville, MD, USA) so as to select for vancomycin-resistant enterococci (VRE). The results documented a statistically significant difference in the ocurrence of low-level VRE such as Enterococcus casseliflavus and Enterococcus gallinarum in vegetarians versus controls. 5 It was suggested that the high rate of low-level VRE carriage in vegetarians could be due to the fact that E. casseliflavus was found to be associated with plants. No association was found between meat consumption and carriage of high-level VRE, as opposed to analyses by another Dutch group. In the same group of vegetarians and controls, lactosepositive Gram-negative (LPGN) bacilli were cultured from the faecal specimens on MacConkey agar. Escherichia coli strains (n 117), identified with the Vitek system, were randomly selected from the LPGN bacilli isolated from faecal samples of 318 vegetarians. As a control group, 101 additional E. coli strains were cultured from the rectal swabs obtained from the 276 controls. Susceptibility to various antimicrobial agents was assessed (see Table) using the disc diffusion method according to the NCCLS. The antimicrobial agents investigated comprised drugs registered for use in humans as well as agents used as growth promoters in animal husbandry. In the absence of accurate NCCLS guidelines for tylosin, zone diameters were defined as following the manufacturer’s criteria: 26 mm, susceptible; 23–25 mm, intermediate; 22 mm, resistant. The enterococci were also screened for high-level gentamicin and streptomycin resistance (MIC 500 mg/L) with Etests (AB Biodisk, Solna, Sweden). The Table displays the results of the susceptibility tests performed for all low-level VRE. When the resistance ratios in vegetarians were compared with controls (Fisher’s exact tests) there appeared to be significantly more bacitracin-intermediate strains in the control group (P 0.0067). The table also provides an inventory of the antimicrobial susceptibility of the E. coli strains. None of the comparisons appears to be significant. However, we do observe a trend towards decreased susceptibility to nitrofurantoin in the control group (P 0.06). Overall, no clear differences seem to exist when the resistance to various antibiotics is assessed in E. coli or enterococci from the two groups. However, when the prevalence of antibiotic-resistant E. coli in community-based vegetarians and volunteers are compared with resistance figures in E. coli from hospitalized patients (Table), the nosocomial strains are markedly more resistant to nearly all of the antibiotics. Apparently, the impact of hospitalization on the prevalence of antibioticresistant bacteria is more important than dietary habits. Our data suggest that the gastrointestinal flora and the prevalence of drug-resistant bacteria vary with dietary habits. This is particularly clear from the enterococcal colonization of the vegetarian gut with vanC enterococci. Interestingly, vanC enterococci from vegetarians are also significantly less susceptible to the antibiotic bacitracin, which has been used in food production animals. We are, however, not aware, of the use of this drug in crop production. Apart from a trend towards a decreased susceptibility Correspondence Journal of Antimicrobial Chemotherapy (2001) 47, 495–502