This report focuses on laboratory stool culturing practices for the variety of vibrios that cause gastroenteritis, particularly V. parahaemolyticus. We did not address the specific tests needed to identify toxigenic V. cholerae O1 or O139. Laboratory diagnosis of Vibrio infections that cause gastroenteritis usually requires isolation of the organism from the stool specimen of a patient with diarrhea. However, many cases of gastroenteritis caused by Vibrio spp. are undetected by clinical laboratories because vibrios are not easily identified on routine enteric media (2). Isolation of vibrios from stool is greatly enhanced through the use of a selective medium, particularly thiosulfate-citrate-bile salts-sucrose agar (TCBS). Although their reliability has not been demonstrated for many Vibrio strains (14), commercial biochemical identification systems are also used for the identification of Vibrio species (18).

In 1989, in response to the increased incidence of Vibrio infections in the states along the Gulf Coast, four Gulf Coast states (Alabama, Florida, Louisiana, and Texas) began systematically reporting infections with Vibrio species to the Centers for Disease Control and Prevention (CDC) (13). To facilitate surveillance for vibrios, the use of TCBS agar for a stool culture was recommended if a patient presented with a compatible diarrheal illness and a history of eating raw seafood (13). However, to our knowledge, the use of TCBS has never been evaluated. In the summer of 1998, during an outbreak of V. parahaemolyticus infections associated with the ingestion of raw oysters in Galveston, Tex. (CDC, unpublished data), we assessed laboratory stool culturing practices for Vibrio species in five Gulf Coast states, where Vibrio infection is reportable. As part of the outbreak investigation, objectives were to ascertain the role of the clinical laboratory in performing surveillance for Vibrio by determining the proportion of clinical laboratories in the Gulf Coast states that routinely cultured stools for Vibrio and the perceived barriers to culturing for Vibrio with TCBS agar.

Among those that received stool specimens for culture, 102 (62%) of 164 laboratories performed at least some stool cultures on-site. Hospital-based laboratories were significantly more likely than independent laboratories to culture at least some stools on-site (86% versus 14%; odds ratio, 3.7; 95% confidence interval, 1.6 to 8.6). Because hospital-based laboratories received more stool specimens and performed more on-site testing than did independent laboratories, 94% (152,797 of 162,559) of all stools received by laboratories in this survey were cultured on-site; the remainder (6%) were sent by the primary clinical laboratory to a reference laboratory. These 102 laboratories reported receiving 88% of stool samples as whole stools and 12% as rectal swabs. Among samples received as whole stools, 18% were submitted with transport media and 82% were submitted without transport media. Among samples received as rectal swabs, 88% were submitted with transport media and 12% were submitted without transport media. The median length of time from stool sample collection to receipt by the laboratory was 2 h. Among 102 supervisors from laboratories that performed on-site stool cultures, 78 (76%) knew that Vibrio infections were reportable in their state. The laboratories managed by these supervisors processed 72% of the stools cultured on-site. Forty-eight (47%) of 102 laboratories that performed on-site stool cultures (representing 63% of stools cultured on-site) routinely sent Vibrio isolates to the state public health laboratory. When asked about the type of patient information to which laboratories routinely had access, only 13 (13%) of 102 laboratories that performed stool cultures reported routinely receiving a patient's food history from the ordering physician. None of the 102 laboratories that processed stool cultures reported changes in their stool culturing practices since 1 January 1998.

Of the 102 laboratories that processed stool cultures, 48 (47%) reported that they used TCBS agar on some or all stool cultures (Table ​(Table1).1). These 48 laboratories processed 85% of the stools that were cultured on-site. Twenty (20%) of 102 laboratories that performed on-site stool cultures used TCBS agar routinely on all stool cultures (which represented 22% of stools cultured in this survey), and 28 (27%) of 102 used TCBS agar on at least some stool cultures (which represented 63% of stools cultured in this survey). One (2%) of 48 laboratories that used TCBS agar reported using TCBS seasonally (May to October). When asked how they confirmed growth of Vibrio on TCBS agar, 40 (83%) of the 48 laboratories that used TCBS agar on some or all stool specimens reported using a commercial biochemical identification system. The 14 Texas laboratories that used TCBS agar on some, but not all, stool cultures were asked what triggered their decision to use TCBS agar. Seven (50%) of the laboratories reported that they used TCBS agar only upon physician request, while the other seven reported using TCBS agar when colony morphology on other enteric agars was suggestive of Vibrio.

The 54 laboratories that did not use TCBS agar on any stool cultures were asked how they screened stool cultures for Vibrio. Thirty-two (59%) of these laboratories reported that they did not screen for vibrios, 20 (37%) used only routine enteric media and a commercial biochemical identification system to identify Vibrio, and 2 (4%) performed a Gram stain. Laboratories that did not use TCBS agar were asked for reasons why they did not use TCBS. Thirty-seven (69%) believed that the local incidence of Vibrio was too low, 23 (43%) thought the cost of testing was too high, and 3 (6%) reported that testing with TCBS agar took too much time.

Cost was perceived by 43% of laboratories to be a barrier to the routine use of TCBS agar in screening for Vibrio spp. Cost per positive culture has been used as a determinant of the practicality of routine laboratory screening for enteric pathogens, such as Escherichia coli O157:H7 (4). Cost per positive culture in this survey was estimated by taking the number of stool samples routinely screened using TCBS during the month of June 1998 in laboratories that met the following criteria: (i) screening all stools for Vibrio with TCBS agar and (ii) isolating a Vibrio organism from TCBS agar. To approximate screening for sporadic Vibrio infections, Texas laboratories were excluded from this cost estimate. The number of stool samples screened during June was multiplied by the cost per TCBS agar plate ($1.50) and then divided by the number of culture-confirmed Vibrio infections identified by these laboratories. Using this formula, cost per positive culture was $72.00. The average treatment cost per culture-positive V. parahaemolyticus infection is $1,000.00 (9). If laboratories were to focus their surveillance for Vibrio by routinely screening all stool samples with TCBS agar during the summer months, more Vibrio infections would be detected, thus reducing cost per positive culture.

Alternatively, laboratories could reduce the cost and make more efficient use of TCBS agar by having access to patient's food histories, particularly during the warmer months. Only 13% of laboratories reported obtaining a food history from the physician. Physicians' knowledge of risk factors associated with various enteric infections is an important clinical diagnostic tool (5, 19). Increased awareness among physicians regarding Vibrio infections may increase the likelihood of obtaining a food history, as well as increase the number of requests for stool cultures to be screened for Vibrio. A food history indicating recent seafood exposure provided to the microbiologist may help those in the laboratory choose which stool specimens to test with TCBS agar, even if a specific request was not made.

The use of commercial systems for the identification of Vibrio species needs further evaluation. Commercial biochemical identification systems such as Vitek (Vitek Systems, Inc., Hazelwood, Mo.) originated in the 1960's. These systems are capable of identifying gram-negative bacteria, gram-positive bacteria, yeasts, and anaerobic organisms. In 1989, the gram-negative identification test kit for the Vitek system was expanded to include members of the Vibrio species (18). In one study, the Vitek system was evaluated for its ability to correctly identify 212 Vibrio isolates, compared with conventional biochemicals and another commercial system, the API20E (Analytab Products, Plainview, N.Y.) (S. Farnham, N. Moss, and J. Scott, Abstr. 89th Annu. Meet. Am. Soc. Microbiol. 1989, abstr. C-263, 1989). The overall correlation of Vitek with the reference identification systems was 95%. Another study used 60 Vibrio isolates to compare three different commercial systems against the use of standard biochemicals; only one, the API20E, was considered to be a valid system for use in the identification of the more commonly isolated members of the family Vibrionaceae (16). Our limited survey indicated that using other enteric agars in conjunction with commercial biochemical identification systems to identify Vibrio spp. yielded accurate results only 50% of the time. However, when TCBS agar was used in conjunction with a commercial system, 88% of isolates were correctly identified. This finding suggests that until further evaluation of a large number of Vibrio strains is completed, TCBS agar should be used for primary laboratory isolation of Vibrio species, with the use of commercial systems for species identification.

Unfortunately, even though Vibrio infections continue to occur in coastal areas (6, 7), the percentage of Gulf Coast laboratories routinely screening stools for Vibrio with TCBS agar was significantly lower (20% versus 30%) than that found in one recent national laboratory survey (20) and approximately the same as that from another national laboratory survey (T. J. Van Gilder, D. Christensen, S. Shallow, T. R. Fiorentino, S. Desai, M. Pass, J. Wicklund, C. Stone, and M. Cassidy, Abstr. 99th Gen. Meet. Am. Soc. Microbiol., abstr. C-419, 1999). Furthermore, the majority (53%) of Gulf Coast laboratories that processed stool specimens on-site never used TCBS agar. Finally, there was no evidence of seasonal use of TCBS agar in our survey. Taken together, these data indicate that TCBS agar is underused in Gulf Coast states.