LETTER

Top Letter References

In the May 1998 issue, Park et al. report their isolation from the stool of a patient with traveler's diarrhea of a nonpathogenic strain of Citrobacter sedlakii which produced positive reactions with various reagents for Escherichia coli O157 antigen (4). These authors have emphasized that additional biochemical testing and/or toxigenicity testing is needed when isolates react with O157 reagents and cite other articles that have described false-positive results on testing for this antigen (1, 5, 7). Since we have evaluated a number of screening protocols for enterohemorrhagic E. coli and E. coli O157, this article provoked an animated discussion in our laboratory, and we offer the following comments.

Sorbitol-MacConkey agar is intended to allow rapid screening for organisms that cannot ferment sorbitol, or that do so very slowly. These plates should be incubated, protected from light, in a non-CO₂ incubator and reviewed within the first 18 to 24 h after incubation. Under these conditions, sorbitol-negative colonies appear colorless. Some sorbitol-negative organisms may lose the colorless appearance when incubated for longer periods of time or in a 5% CO₂ atmosphere. Park et al. indicate that the C. sedlakii isolate in question was sorbitol positive and produced pink colonies on sorbitol-MacConkey agar, not having the colorless appearance that would have been observed for sorbitol-negative organisms. Most protocols that use sorbitol-MacConkey screening do not call for further evaluation of sorbitol fermenters.

Even when sorbitol-negative isolates are recovered, it is prudent to identify these isolates biochemically before testing for the O157 antigen. Other sorbitol-negative, gram-negative organisms that may agglutinate E. coli O157 latex reagents include Alcaligenes faecalis and Morganella morganii (6), as well as some sorbitol-negative members of the genus Escherichia other than E. coli (5). Similar protocols for grouping of Salmonella, Shigella, and Streptococcus isolates require identification by conventional means before serologic testing is performed. The necessity to identify an isolate by conventional means before antigenic testing to avoid possible false-positive reactions has been observed repeatedly (2, 8). The fourth edition of the Manual of Clinical Microbiology states: "Because of the possibility of serological cross-reactions, it is essential that isolates be subjected to rigorous biochemical testing before serological analysis is performed ... for identification of Salmonella and Shigella species and certain E. coli isolates (2). Although the identifications Park et al. obtained for their isolate were not consistent among the commercial or conventional systems used, no system identified the isolate as an Escherichia species. No information is presented to indicate that their patient's clinical course was unusual, or that complications occurred to prompt continued evaluation. Any concerns that remained regarding possible Shiga-like toxin production by an atypical isolate should have been assuaged by the negative result obtained with the Premier EHEC test (Meridian Diagnostics, Cincinnati, Ohio).

In a recent review of diarrheagenic E. coli, Nataro and Kaper noted that the failure to accurately detect E. coli O157 has led to unnecessary procedures, including exploratory surgery, hemicolectomy, colonoscopies, barium enemas, and appendectomies (3). We must also consider that false-positive laboratory reports can result in delayed consideration of alternative diagnoses or therapeutic interventions in the acutely ill patient, or in unnecessary anxiety regarding possible sequelae. This report illustrates the importance of following standard procedures in the clinical laboratory to maximize the likelihood of a meaningful laboratory result, even when circumstances or good intentions tempt us to do otherwise.

Ed. Note: The authors of the published article did not feel that a response was necessary.