Stenotrophomonas maltophilia in salad.

To the Editor: Stenotrophomonas maltophilia has emerged as an important nosocomial pathogen, especially in debilitated and immunocompromised persons (1). However, comparatively little is known of the epidemiology of this bacterium, and sources and routes of transmission of S. maltophilia are not well understood. The bacterium is widely distributed in the environment, including in plant rhizospheres (2). Although environmental sources such as ice-making machines have been implicated in outbreaks of nosocomial S. maltophilia 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-imipenem-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 vancomycin-resistant Enterococcus faecium, 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, minocycline, 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 × 102 to 1.96 × 105 CFU/g (mean 1.75 × 105 CFU/g, median 7.05 × 103 CFU/g). All isolates were susceptible to ciprofloxacin, colistin sulfate, minocycline, and trimethoprim-sulfamethoxazole. Eight (89%) were resistant to chloramphenicol; 7 (78%) to piperacillin/tazobactam; 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 increase before use.

All of the 9 isolates examined exhibited resistance to >2 of the 8 agents tested, with 2 isolates resistant to 4 compounds and thus had resistance phenotypes similar to those of strains associated with human infection. These findings are in agreement with those of Berg (2), who reported multiple resistances among isolates of S. maltophilia associated with oilseed rape. However, as with most strains of clinical origin, all the isolates tested here remained susceptible to trimethoprim-sulfamethoxazole and minocycline.

Although other investigators have examined salad products for the prevalence of pathogenic bacteria (4), they did not attempt to isolate S. maltophilia. Another study may have underestimated the prevalence of the bacterium in these products since a selective medium was not used (5). We have shown in a clinical setting that the use of a medium selective for S. maltophilia improves the recovery of this bacterium (6). Furthermore, the optimal growth temperature of S. maltophilia is 30°C, and incubation at higher temperatures (7) may reduce the likelihood of recovering the bacterium from food products. We recommend that future studies of S. maltophilia in food products use a medium such as VIA and that cultures be incubated at 30°C.

Prepackaged, ready-to-eat salads, such as those examined in this study, are washed in chlorinated water before sale. This measure is clearly insufficient to remove S. maltophilia from these items, possibly because the bacterium may exist in biofilms in some of the components of these products. S. maltophilia is capable of forming biofilms on a number of materials (8). Alternatively, products may become contaminated from environmental sources in production plants after washing.

The importance of S. maltophilia in ready-to-eat salads, which are marketed in a manner that assumes the product does not need washing before consumption, is unknown; nevertheless the presence of the bacterium in these products serves to highlight recommendations that these items should be avoided by severely immunocompromised persons, especially those with neutropenia (9). Recently, Apisarnthanarak et al. (10), in a prospective study of hospitalized oncology patients, identified intestinal colonization with S. maltophilia in 4 (9.5%) of 41 patients, which emphasizes that foodstuffs may be a potential source of this bacterium for some patients. This is a preliminary study, however, and further studies are needed, in particular, molecular typing of food and human-associated isolates, to investigate the hypothesis that intestinal carriage of S. maltophilia may follow consumption of contaminated foodstuffs.