In October 2009, the CDC published guidelines for clinical laboratories recommending that all stools submitted for diagnosis of acute community-acquired diarrhea be simultaneously cultured as well as tested for Shiga toxin. Routine use of selective and differential agar to detect Escherichia coli
O157:H7 and either immunoassays or molecular methods to detect the presence of Shiga toxin is the currently recommended laboratory testing strategy for Shiga toxin-producing Escherichia coli
). According to published reports, approximately 96,534 STEC O157 and 168,698 non-O157 infections occur each year. These infections are associated with more than 3,600 hospitalizations and 30 deaths annually (2
). However, the national isolation rate of STEC O157 in 2009 was 0.72 per 100,000 population and the isolation rate of STEC non-O157 was 0.29 per 100,000 population, suggesting that there is a variation in the rate of non-O157 occurrence around the country (Centers for Disease Control and Prevention, Shiga toxin-producing Escherichia coli [STEC] annual summary, 2009 [http://www.cdc.gov/ncezid/dfwed/PDFs/national-stec-surv-summ-2009-508c.pdf
, accessed 11 November 2012]). Overall, STEC is detected in approximately 0% to 4.1% of stools submitted for testing at clinical laboratories; these percentages rival those for other bacterial pathogens (1
). Because signs and symptoms of STEC infection can vary significantly among individuals before the development of life-threatening complications, the CDC's current recommendations of universal testing were implemented, with the goals of prompt detection of outbreaks, early diagnosis for improved patient care, identification of additional STEC infections, and detection of all STEC serotypes, including O157 and non-O157 serotypes.
Beginning in January 2008, the clinical microbiology laboratory at our institution, an 800-bed tertiary medical center and children's hospital, implemented a universal STEC screening protocol by adding Shiga toxin antigen testing as part of the battery of tests performed in each stool sample submitted for bacterial culture. From January 2008 to December 2011, a total of 5,017 consecutive fecal samples were received and cultured to identify bacteria positive for STEC serotypes and the presence of Shiga toxins. Samples were simultaneously cultured and tested for Shiga toxins 1 and 2. Culture was achieved according to routine protocols using a sorbitol-MacConkey (SMAC) plate to detect sorbitol-negative colonies. The GN enrichment broth was prepared in-house from dehydrated medium (Becton Dickinson, Sparks, MD). The 8-ml broth tube was inoculated with either 50 μl of unpreserved specimen or a 3- to 4-mm round pellet of stool. The Shiga toxin detection was performed with the Immunocard STAT! enterohemorrhagic E. coli
(EHEC) test (Meridian Bioscience, Cincinnati, OH) using a 150-μl aliquot from the enrichment broth that had been incubated for 16 to 24 h as recommended by the manufacturer. The Immunocard STAT! EHEC test is an immunochromatographic rapid test that utilizes monoclonal antibodies labeled with red-colored gold particles and detects and differentiates Shiga toxin 1 and 2 antigens present in the sample (3
). SMAC culture plates were examined for sorbitol-negative colonies, and, if colonies were present, 4 or 5 colonies were tested for the O157:H7 serotype using latex agglutination (Remel Products, Thermo Fisher Scientific, Lenexa, KS). Specimens were considered positive or negative for STEC according to the following criteria. (i) If Shiga toxin immunochromatographic assay (IA) detected the presence of Shiga toxin 1 or 2 or both on enrichment broth and sorbitol-negative E. coli
was recovered by culture and confirmed as O157, the specimen was deemed positive for STEC by Shiga toxin IA and by culture. The organism was forwarded to the North Carolina State Laboratory of Public Health (NCSL). (ii) If the specimen was Shiga toxin positive on enrichment broth and sorbitol-positive E. coli
was recovered by culture, the Shiga toxin test was performed on the colony. If Shiga toxin was detected from the colony, the organism was forwarded to NCSL for serotyping. If an STEC serotype was identified by NCSL, the specimen was considered positive by Shiga toxin IA and by culture. (iii) If the specimen was Shiga toxin positive but no E. coli
was recovered in culture, the enrichment broth was subcultured onto a SMAC plate. If no E. coli
was detected in the broth subculture SMAC plate either, the specimen was deemed negative by culture but positive by Shiga toxin IA. (iv) If the enrichment broth was Shiga toxin negative and culture grew sorbitol-negative E. coli
O157, the Shiga toxin test was performed from the organism's colony to confirm that the organism was a Shiga toxin producer. If Shiga toxin was detected from the organism, the specimen was considered negative by IA on enrichment broth and positive by culture. (v) If the specimen was Shiga toxin negative on enrichment broth and culture grew sorbitol-positive E. coli
, no further testing was done and the specimen was deemed negative for STEC by Shiga toxin and culture testing.
Over 4 years, a total of 12 samples were positive for ETEC either by Shiga toxin testing, culture, or both (). The rate of isolation was estimated using the number of specimens in which STEC was detected, either by culture or IA, divided by the total number of stool specimens tested each year and was found to be fairly constant, with rates of 0.3%, 0.2%, 0.2%, and 0.3% from 2008 to 2011, respectively. Among the 12 STEC cases detected, patient ages ranged from 2 to 68 years, with 6 ≤10 years of age, 4 with ages between 15 and 43 years, and 2 >60 years of age. Gender demographics included 6 male and 6 female patients.
Laboratory findings in 12 STEC cases detected from 2008 to 2011
Overall, 8 cases were detected by both culture and Shiga toxin IA, 3 were detected by culture only, and 1 by IA only (). Among the 8 specimens Shiga toxin positive by IA and culture, 6 grew sorbitol-negative E. coli confirmed as O157:H7 serotype and 2 grew sorbitol-positive E. coli confirmed as non-O157 serotype (both O103) and were positive for Shiga toxin 1 only. All 3 cases detected by culture grew only sorbitol-negative E. coli O157. Although Shiga toxin testing was not repeated from the enrichment broth, Shiga toxin testing was performed on these isolates growing on the SMAC culture plate on the day of isolation, and they were all positive by Shiga toxin IA. The NCSL confirmed that these isolates were serotype O157, and therefore these three cases were considered positive for STEC by culture. One specimen was Shiga toxin positive with no E. coli organisms isolated from the primary culture or from the subculture performed from the enrichment broth. Although the presence of STEC could not be confirmed by culture, this specimen was included in the 12 cases of STEC described in this study. The 2 serotype O103 isolates were isolated 6 months apart in 2011 and were likely unrelated epidemiologically. One was isolated from a 10-year-old patient and the other from a 62-year-old patient. Among the 9 Shiga toxin-positive specimens, 2 were positive for Shiga toxins 1 and 2, 3 were positive for Shiga toxin 1 only, and 4 were positive for Shiga toxin 2 only. A total of 4 hemolytic-uremic syndrome (HUS) incidents were identified in our study—one was detected by culture alone and three were detected by culture and Shiga toxin IA. Among the four HUS incidents, three were seen in pediatric patients and were associated with the E. coli O157:H7 serotype and one was seen in an adult patient and was associated with an E. coli O103 serotype.
Our results demonstrated an overall STEC isolation rate of 0.24% (0.2% for O157 and 0.04% for non-O157), which is lower than the national isolation rate of 0.72 per 100,000 population reported in 2009 (CDC [http://www.cdc.gov/ncezid/dfwed/PDFs/national-stec-surv-summ-2009-508c.pdf
]) but closer to the isolation rate range of 0.26 to 1 per 100,000 population reported in North Carolina in 2009 (2
), to 0.13% in the southeast region of the United States from 1990 to 1992 (4
), and to the 0.3% national annual rate from 1995 to 2000 (5
). Indeed, isolation rates differ considerably by state: states in the upper Midwest have the highest STEC O157 isolation rates (range, 1.01 to 5.3%), with much lower rates in states in the South (range, 0.0 to 1.0 per 100,000 population) (2
). Our data reflect a low STEC prevalence similar to that from Upstate Medical University Hospital, NY, between 2007 and 2008 (0.09%). This hospital, a 400-bed medical center and children's hospital, utilized the ProSpecT STEC microplate assay for the detection of Shiga toxin (Remel Inc., Lenexa, KS) (6
Our study's findings are consistent with the assessment that testing by Shiga toxin IA has important inherent limitations for the detection of STEC. First, false positives are not uncommon and therefore positive Shiga toxin testing results must be confirmed with culture. Two separate cases of gastroenteritis during 2005, one in New York and one in North Carolina, highlight the importance of culture confirmation of STEC after a positive Shiga toxin result utilizing immunoassays (7
). In our study, we found one possible false positive, a sample positive for Shiga toxin 1 by IA but from which E. coli
was not recovered from the primary culture or from the subculture from broth. However, this case may still represent clinical disease. Second, false negatives also appear to be a problem. In our study we had 3 possible false negatives by Shiga toxin IA. All of these grew E. coli
O157, with moderate to heavy growth on SMAC. Similar findings have been reported by other investigators (8
). Shiga toxin testing was performed from the sorbitol-negative colonies growing on these cultures, and the test was positive for Shiga toxins 1 and 2 in one specimen and for Shiga toxin 2 in the other two cases. These cases would have been missed if only Shiga toxin testing was performed from the enrichment broth with no simultaneous culture. The reason for this discrepancy (negative Shiga toxin result from the enrichment broth but culture with sorbitol-negative E. coli
O157 confirmed as a Shiga toxin producer) is not clear but could have been due to technical errors, presence of inhibitors in the specimens, or the concentration of the toxin being below the sensitivity of the assay when it was performed from the broth (1
With respect to testing by culture, detection of Shiga toxin replicated that of culture in all but one case—the possible false positive discussed above. Moreover, Shiga toxin testing failed to detect 3 positive samples on the first attempt; the samples were determined to be positive by culture. However, it must be noted that Shiga toxin IA helped to detect two sorbitol-positive STEC (both serotype O103) specimens by triggering a petition to NCSL for confirmation of serotype. At the same time, it should be noted that both of these cases involved individuals from high-risk groups. In this way, our study found that universal simultaneous testing by Shiga toxin IA and culture yielded little benefit over testing by culture alone and no demonstrable benefit over simultaneous testing of Shiga toxin IA and culture restricted only to high-risk groups for the detection of STEC. In contrast, however, one study at 2 Seattle sites, including the Children's Hospital and Regional Medical Center Emergency Department and a private pediatric practice, found that sole reliance on Shiga toxin testing by enzyme immunoassay (EIA) (Premier EHEC; Meridian Bioscience, Cincinnati, OH) would have underdetected E. coli
O157:H7 specimens by 11% and that sole reliance on SMAC agar screening would have underdetected the STEC specimens by 28% (9
). An explanation for this discrepancy with our study could be a difference in the sensitivities of the Shiga toxin assays used and different patient populations.
In our study, HUS was detected in 4 individuals. It is important to note that one of these cases, the case in which STEC serotype O103 producing only Shiga toxin 1 was isolated from an adult patient, was detected because the Shiga toxin IA from the enrichment broth was positive, and this finding prompted us to forward the sorbitol-positive E. coli growing on the SMAC culture to the NCSL for serotyping. This case would have been missed if only culture had been performed because the organism grew as sorbitol positive and could have been considered non-STEC. Nevertheless, it must be noted that all cases of HUS, including this case, were screened under circumstances of preexisting clinical suspicions for HUS before the detection of positive cultures. This suggests that universal testing may not hold an advantage over selective testing for patient management, undermining the rationale behind the current CDC recommendation for simultaneous testing by Shiga toxin and culture, at least for areas of low prevalence.
The laboratory protocol that we developed based on the recommendations of CDC for simultaneous testing by Shiga toxin IA and culture is very cumbersome to implement and can be very complicated to interpret due to discrepancies between the two tests and the need for additional retesting to evaluate these discrepancies. In a clinical setting these constitute serious drawbacks. Furthermore, Shiga toxin testing adds a high cost to the stool work-up. In our institution, the cost of the stool culture battery of tests increased drastically when Shiga toxin testing by IA was added to each sample. This represented a significant budget burden over the 4-year period.
Finally, we note other findings of our study, not dealing strictly with a comparison between approaches to STEC detection. Investigators have reported that the highest isolation rates for both O157 and non-O157 STEC were in children <4 years old (CDC [http://www.cdc.gov/ncezid/dfwed/PDFs/national-stec-surv-summ-2009-508c.pdf
]). In our study, only one patient with O157 STEC was <4 years old while 4 patients with O157 and 1 with O103 had ages between 4 and 10 years.
There were a total of 5 patients (18%), compared to our 4 patients (33%), who had the diagnosis of HUS. Shiga toxin 2 was found in almost all E. coli
O157:H7 specimens, which is similar to our findings, with Shiga toxins 1 and 2 or 2 only detected in 6 cases culture positive for the O157 serotype. Non-O157:H7 specimens were not found to be positive for Shiga toxin 2. This is again similar to the findings of our study, where the two non-O157 isolates recovered were positive for Shiga toxin 1 only (9
A recent publication demonstrates that, although the number of laboratories testing for Shiga toxin has increased, only 2% of laboratories reported simultaneous culture for O157 STEC and Shiga toxin testing (10
). This study presents data from 4 years of simultaneous testing in a routine clinical setting and shows that simultaneous testing can be very complicated to interpret due to discrepancies between the two tests and the need for additional testing to evaluate these discrepancies. The findings of this study could be used as a baseline for the comparison of other approaches which measure the cost-effectiveness of different methods of detection of STEC in stool samples which are potentially more suitable for routine testing in clinical laboratories. In the future, molecular testing may offer a more specific and less time-consuming method for STEC detection than simultaneous culture and Shiga toxin detection in stool samples (11
Based on our 4-year experience with universal STEC screening in a clinical setting, universal simultaneous testing by Shiga toxin IA and culture has been shown to yield little benefit over testing by culture alone and no demonstrable benefit over simultaneous testing restricted to high-risk groups for either the detection of STEC or for patient care. In contrast, the addition of Shiga toxin IA testing has significant drawbacks in effort, time, and cost. To improve the cost-benefit analysis result for STEC testing in areas of low prevalence of non-O157:H7 E. coli, simultaneous culture and Shiga toxin testing restricted to high-risk groups may be preferable to universal testing.