Stenotrophomonas maltophilia in Salad

To the Editor: Stenotrophomonas maltophilia has emerged as an important nosocomial pathogen, especially in debilitated and immunocompro-mised persons (1). However, comparatively little is known of the epidemiology of this bacterium, and sources and routes of transmission of S. mal-tophilia are not well understood. The bacterium is widely distributed in the environment, including in plant rhi-zospheres (2). Although environmental sources such as ice-making machines have been implicated in outbreaks of nosocomial S. maltophil-ia sepsis, the source of infection in other outbreaks and in sporadic cases often remains unidentified. Food as a source of the bacterium has not been investigated; given the association of the bacterium with plants, we investigated the prevalence of the bacterium in salad vegetables. Salads were purchased from reputable supermarkets and transported immediately to the laboratory. Ten grams of salad was homogenized in 90-mL sterile saline in a stomacher for 1 to 2 min. Aliquots (200 µL) of decimal dilutions of the homogenate were plated onto vancomycin-imipen-em-amphotericin B (VIA) agar (3) and incubated for 24 to 48 h at 30°C. Occasional imipenem-resistant environmental bacteria, such as Janthinobacterium lividum or van-comycin-resistant Enterococcus faeci-um, may grow on VIA, but the medium contains a mannitol/bromothymol blue indicator system, allowing these bacteria, which produce acid from mannitol, to be distinguished from S. maltophilia. Putative S. maltophilia colonies were further identified by the API 20NE system (bioMérieux, Marcy l'Etoile, France). Susceptibilities of 9 confirmed isolates to ceftazidime, chloramphenicol, colistin sulfate, gentamicin, minocy-cline, piperacillin/tazobactam, and trimethoprim-sulfamethoxazole were determined by using a disk diffusion method. Pseudomonas aeruginosa (NCTC10662) was used as a control. Because disk diffusion is not a reliable method for determining the susceptibility of S. maltophilia to quinolone antimicrobial agents (1), the Etest was used. S. maltophilia was cultured from 14 (78%) of 18 salads. Numbers ranged from 1.50 × 10 2 to 1.96 × 10 5 CFU/g (mean 1.75 × 10 5 CFU/g, median 7.05 × 10 3 CFU/g). All isolates were susceptible to ciprofloxacin, col-istin sulfate, minocycline, and trimethoprim-sulfamethoxazole. Eight (89%) were resistant to chlorampheni-col; 7 (78%) to piperacillin/tazobac-tam; 5 (56%) to ceftazidime, and 2 (22%) to gentamicin. All products examined were labeled " washed and ready to eat, " and, thus, consumers would be unlikely to wash these products before consumption. The growth characteristics of S. maltophilia in products of this type, especially if subject to temperature abuse, are unknown, but in the domestic setting, numbers of the bacterium may …