The Emerging Clinical Importance of Non-O157 Shiga Toxin Producing Escherichia coli

INVITED ARTICLE EMERGING INFECTIONS James M. Hughes and Mary E. Wilson, Section Editors The Emerging Clinical Importance of Non-O157 Shiga Toxin Producing Escherichia coli Kristine E. Johnson, 1 Cheleste M. Thorpe, 2 and Cynthia L. Sears 1 1 Division of Infectious Diseases, Johns Hopkins University School of Medicine, Baltimore, Maryland, and 2 Division of Infectious Diseases and Geographic Medicine, Tufts University School of Medicine, Boston, Massachusetts In 1982, hemorrhagic colitis and hemolytic-uremic syndrome were linked to infection with Escherichia coli O157:H7, a serotype now classified as Shiga toxin producing E. coli (STEC). Thereafter, hemorrhagic colitis and hemolytic-uremic syndrome associated with non-o157 STEC serogroups were reported, with the frequency of non-o157 STEC illness rivaling that of O157:H7 in certain geographic regions. In the United States, non-o157 E. coli may account for up to 20% 50% of all STEC infections. A high index of suspicion, paired with options to test for non-o157 STEC infection, are necessary for early recognition and appropriate treatment of these infections. Supportive care without the use of antibiotics is currently considered to be optimal treatment for all STEC infections. This commentary provides a perspective on the non-o157 STEC as human pathogens, how and when the clinician should approach the diagnosis of these organisms, and the challenges ahead. In 1982, infection with Shiga toxin producing Escherichia coli (STEC) O157:H7 strains was linked to hemorrhagic colitis and the hemolytic-uremic syndrome (HUS), defining a new foodborne zoonosis. Since then, 250 different O serogroups of E. coli have been shown to produce Shiga toxin, and 1100 of these STEC have been associated with sporadic and epidemic human l disease. Although the virulence of E. coli O157:H7 is well described, over the past 20 years, the capacity of the heterogeneous family of non-o157 STEC to induce human disease has spurred considerable, even heated, debate [1 3]. The data, although presented in variable detail, indicate that some non- O157 STEC unequivocally cause human disease and likely account for 20% 50% of STEC infections (an estimated 37,000 cases of illness) annually in the United States [4]. This commentary collates selected data to provide a perspective on the non-o157 STEC as human pathogens, how and when the clinician should approach diagnosis of these organisms, and the challenges ahead. We focused on reports that included 5 patients infected with non-o157 STEC and that provided clinical details of the disease spectrum observed. Received 2 March 2006; accepted 29 August 2006; electronically published 9 November 2006. Reprints or correspondence: Dr. Kristine E. Johnson, Div. of Infectious Diseases, Johns Hopkins University School of Medicine, 1830 E. Monument St., Rm. 401, Baltimore, MD 21205-2100 (kjohns87@jhmi.edu). Clinical Infectious Diseases 2006; 43:1587 95 2006 by the Infectious Diseases Society of America. All rights reserved. 1058-4838/2006/4312-0011$15.00 NON-O157 STEC CLINICAL DISEASE: SUBTLE TO DEADLY Non-O157 STEC pose a substantial dilemma for the clinician, because these bacteria cause illnesses indistinguishable from O157-induced disease and many other foodborne enteric infections. Data defining the overall frequency of non-o157 STEC compared with that of other enteric pathogens are limited. In this review, we analyzed nonoutbreak studies to try to define the community base of non-o157 STEC associated disease. In 2 studies involving 11 sites in the United States in 1997 and 4 sites during the period 1998 1999, screening of 13,798 stool samples by EIA for STEC yielded a frequency of 0.9% (120 cases of STEC) [5]. Of these 120 cases, isolates were available from 104; 48 (46%) of the isolates were non-o157 STEC, and 56 (54%) were O157 STEC. These isolation rates are similar to rates for Shigella species in some locales. Such data suggest that there is a significant burden of non-o157 STEC associated disease in the United States that is underappreciated. Further, global hot spots exist in which non-o157 STEC serogroups dominate over O157 serogroups, including Argentina, Australia, and Germany (table 1). For example, in Germany, non- O157 STEC account for up to 80% of STEC-associated l illnesses [14]. Clinical manifestations of non-o157 STEC infection range from mild, watery to hemorrhagic colitis, HUS, and death; illness due to non-0157 STEC may be equivalent in severity to illnesses induced by E. coli O157:H7. The clinical EMERGING INFECTIONS CID 2006:43 (15 December) 1587

Table 1. Global identification of non-o157 Shiga toxin producing Escherichia coli (STEC) serogroups. Country Common non-o157 STEC serogroups identified a Patients with non-o157 STEC infection among patients with STEC infection, b % No. of cases of non-o157 STEC infection/ no. of cases of STEC infection United States O26, O45, O103, O104, O111, O119, O121, O145, OR 44 c 62/141 [6 8] Canada O2, O26, O91, O103, O111, O113, O145 20 69/341 [9 12] United Kingdom O26, O55, O145 28 15/53 [13] Germany O26, O55, O76, O91, O103, O111, O113, O118, O128, O145, O146, ONT, OR Reference(s) 44 c 327/748 [1, 14 16] Spain O26, O98, O118, O145, O150 78 87/111 [15, 17] Italy O26, O55, O103, O111, O145 34 d 92/264 [18, 19] Czech Republic O1, O5, O26, O55, O111 57 16/28 [20, 21] Belgium O26, O91, O111, O113, O118, OR, ONT 63 41/65 [22] France O91, O103, O113, OX3 33 e 39/118 [23 25] Denmark O26, O103, O111, O117, O121, O128, O145, O146, O174, OR 75 493/655 [26, 27] Finland OR, O15, O26, O103, O145, O174 53 f 57/107 [28 31] Sweden O8, O117, O121, OR, ON 33 20/60 [32] Australia O6, O26, O48, O98, O111 69 31/45 [33, 34] New Zealand O22, O26, ONT, OR 100 16/16 [35] Chile O26, O111 63 5/8 [36] Argentina O8, O26, O103, O113, O121, O145, O174, ONT 40 38/96 g [37] Japan O26, O111, O128, O145, O165 19 h 10/52 [38, 39] NOTE. Serogroups were identified via isolation from stool culture and/or serological evidence of STEC infection in the context of clinical illness. Only serogroups reported twice or more across studies evaluated are reported. Only studies evaluating STEC infection due to both O157 and non-o157 serogroups wereconsidered. Data are presented as a percentage of case patients with non-o157 STEC infection of all STEC infections identified in published reports from each country. When identified, domestic strains alone are included in the totals. When reported, number of non-o157 STEC infections includes those with stool culture isolate, serological evidence of infection, or both. Data are limited by missing and unreported information internationally, including reports of STEC infection in which no serogroup identification was performed. Asymptomatic STEC excretors were excluded from the analysis when identified. ONT, O nontypeable; OR, O rough. a Non-O157 infection was defined as (1) absence of O157:H7 on stool culture and/or absence of circulating antibodies to O157 lipopolysaccharide with fecal cytotoxin assays neutralizable by antiserum to Shiga toxin, monoclonal antibody for Shiga toxin 2, or antibodies to Shiga toxin 1 and Shiga toxin 2; or (2) stx1 and/or stx2 DNA probe positive stool isolates; or (3) a 3-fold increase in serum antibody titers to Stx1 or Stx2 without serological evidence of O157:H7 or isolation of O157:H7 on stool culture. If only a subset of total STEC cases reported was identified by serogroup, this subset was used as the denominator for all STEC reported. Serogroups are included if 2 strains of a serogroup were reported. b When identified, cases with dual infection of non-o157 with O157 STEC were not included in the data reported. c Outbreak and single-case reports are excluded from this calculation. d Non-0157 diagnosed by serological testing alone in a subset of cases. e Because of limited information in the studies available, only [25] was used for the calculation. f For calculation, data are from reference [30] only. g Denominator represents the total number of strains isolated after correction for duplicate subjects. h Outbreak cases were not included in calculation of the non-o157 isolation rate. variability of disease noted with non-o157 STEC may be related to their notable genetic heterogeneity. At least 15 studies help to define the breadth of non-o157 STEC related illness [1, 6, 9, 13 15, 17, 26, 32, 37, 40 44] (table 2). Although non-o157 serogroups appear to cause watery l illness more often than bloody l illnesses [6, 45], it is equally apparent that bloody associated with STEC infection cannot be presumed to be due to O157 infection. Conversely, O157 infection does not inevitably cause hemorrhagic colitis. A 2-year prospective study of 5415 cases of sporadic in Canada revealed significant differences between the clinical syndromes associated with non-o157 and O157 STEC infection [9]. Patients with non-o157 STEC infection had a longer duration of (mean duration, 9.1 days vs. 5.7 days; P!.001) and less frequent bloody (42% vs. 97%; P!.001). In contrast, the rate of abdominal pain, vomiting, and fever of 138 C did not differ between patients with non-o157 infections and those with O157 infections. In a recent study of 343 Danish patients with STEC infections [26], serogroups O157 and O103 were independent risk factors for bloody on multivariate analysis with indistinguishable ORs and 95% CIs, confirming that certain non-o157 serogroups can be as virulent as O157:H7 strains. The virulence of non-o157 strains is further illustrated by a study of 394 children with HUS in Germany and Austria, in which 57% of children with non-o157 illness and 71% of those with O157 illness presented with bloody. Although statistically significant ( P!.05), this difference is clinically meaningless [15]. 1588 CID 2006:43 (15 December) EMERGING INFECTIONS

Table 2. Clinical manifestations of non-outbreak disease due to non-o157 Shiga toxin producing Escherichia coli (STEC). Reference Year of publication Design (dates) Population Location Non-O157 STEC infections identified/ no. of patients evaluated a Clinical manifestations, no. of patients with non-o157 associated disease/total no. of patients (%) Nonbloody Bloody HUS Strengths and/or limitations Pai et al. [9] 1988 Prospective study (1984 1986) Lopez et al. [42] 1989 Prospective study (1986 1988) Kleanthous et al. [13] 1990 Prospective study (1985 1988) Bitzan et al. [40] 1991 Retrospective study (1985 1989) Bockemuhl et al. [41] 1992 Retrospective study (1987 1990) Bielaszewska et al. [44] 1994 Prospective study (1988 1991) Huppertz et al. [14] 1996 Prospective study (1994) Hospitalized children and adolescents!16 years old with Beutin et al. [1] 1998 Retrospective multicenter study (1996) Gerber et al. [15] 2002 Prospective, multicenter study (1997 2000) Jelacic et al. [6] 2003 Prospective study (1988 2000) Wielender-Olsson et al. [32] 2002 Retrospective study (1997 1999) Blanco et al. [17] 2004 Prospective study (1992 1999) Ethelberg et al. [26] 2004 Prospective study (1997 2003) Brooks et al. [43] 2005 Prospective sample (1993 2004) Rivas et al. [37] 2006 Prospective study (2001 2002) Children and adults with Canada 29/5415 18/19 (95) 8/19 (42) 2/19 (11) Clinical comparison of O157:H7-associated cases with non-o157 associated cases Children with and/or HUS Argentina 62/95 9 (15) NS b 14 (23) Specific non-o157 serogroups not identified Children with HUS United Kingdom 15/196 7 (47) 8 (53) 15 (100) Clinical data provided for each case Children and adults with mucous or bloody Germany 15/155 13 (87) 6 (40) 6 (40) Clinical data provided for each case Children and adults with Germany 49/62 27 (55) 8 (16) 18 (37) Clinical data limited, retrospective design Children with Czech Republic 23/1239 6 (26) 6 (26) 10 (43) Clinical data limited Non-O157 isolates from adults and children submitted to a central laboratory Hospitalized children and adolescents!16 years old with Adults and children with clinical indication for stool culture Adults and children with clinical indication for stool culture Adults and children with clinical indication for stool culture Germany 11/468 11 (100) 2 (18) 1 (9) Well-characterized cohort Germany 83 c,d,e 72 (87) 7 (8) 4 (5) Well-characterized cohort Austria and Germany 186/394 c 81/87 (93) 47/83 (57) 90 (48) HUS studied only; clinical comparison of O157:H7-associated with non- O157 associated HUS cases United States 50/6300 NR 21/37 (57) NR Clinical data limited Sweden 20/5617 7 (35) 9 (45) 2 (10) Well-characterized cohort Spain 87/5054 f 59 (68) 18 (21) 1 (1) Clinical comparison of O157:H7-associated with non-o157 associated HUS cases Adults and children with STEC infection Denmark 262/343 c NR 69 (26) 10 (4) Well-characterized cohort Non-O157 isolates from adults and children submitted to CDC United States 940 d,g NR 75/292 (26) 21/292 (7) Clinical data available from only 31% of source patients Children with and/or HUS Argentina 38/99 9 (24) 28 (74) 5 (13) Well-characterized cohort NOTE. Studies are ordered chronologically by publication year. Non-O157 STEC infection was identified by fecal culture and/or a positive result of any of the following tests: cytotoxin neutralizing assay, stx DNA probe assay, or serology. Unless noted, is non-bloody. Cases of may have been associated with HUS; there is overlap in the cases reported for each clinical category, such that the syndromes of (non-bloody vs. bloody) and HUS may have been noted in a single patient. Studies were selected for systematic analysis of STEC strains and reporting of clinical features to demonstrate the spectrum of non-o157 STEC illness. This is intended to provide a global perspective on the diversity of clinical data and serogroup identification. CDC, Centers for Disease Control and Prevention; HUS, hemolytic-uremic syndrome; NR, not reported; NS, not specified. a When HUS was studied prospectively, data reported are total number of HUS cases/total number of non-o157 STEC infections identified. Asymptomatic cases of infection due to non-o157 STEC were eliminated from the analysis when symptom-free data were provided. c b Bloody was reported only as a percentage of all reported illnesses; thus, we were unable to derive non-o157 associated bloody. Number of STEC strains identified, excluding dual infection with STEC of 2 different serogroups and any asymptomatic cases where applicable. d Study of non-o157 associated disease only. e Total number of non-o157 strains reported is 126. Serogroup identification was reported for 87. Serogroup and clinical information were reported for 83. Total number of non-0157 STEC strains is 87. Serogroup and clinical information were reported for 85. g Clinical information was available for a limited number of patients with non-o157 STEC infection. Serogroups reported as other are considered to be non-o157 STEC. f

HUS is a triad of acute renal failure, microangiopathic hemolytic anemia, and thrombocytopenia. In the majority of cases in children, HUS occurs after a l illness, but bloody is not a prerequisite. Worldwide, -associated HUS largely accounts for morbidity and mortality due to STEC, resulting in death (in up to 5%) and frequent permanent renal injury (at an estimated rate of 25%) [46]. Because most of the data detailing the risk of progression to HUS, HUS-related mortality, and development of permanent renal injury following HUS have been obtained from O157:H7-associated l outbreaks, it is unknown whether these data also will be true in l illness associated with non-o157 strains. However, it is clear that infection with non-o157 strains can result in HUS. In a 4-year study of children and adults with HUS in the United States, 5 (16.7%) of 30 patients had infection due to non-o157 strains [7]. Argentina has the highest global rates of HUS, estimated annually at 12.2 cases per 100,000 [37] among children younger than 5 years, and non-o157 strains account for nearly all of this disease burden [42]. Similar findings are reported from Australia [33]. In Germany and Austria, a prospective evaluation of sporadic cases of HUS in children revealed non-o157 STEC in nearly one-half (43%) of patients studied [47]. Additional studies from Germany, Belgium, Finland, the Czech Republic, and Italy also suggest that the rate of isolation of non-o157 strains exceeds that of O157 strains among patients with and HUS in these regions. Certain non-o157 serotypes, particularly O111, trigger sporadic and outbreak-associated HUS more frequently than other non-o157 STEC. In a recent sampling of 940 non-o157 human isolates sent to the Centers for Disease Control and Prevention (Atlanta, GA) by public health laboratories in 42 states and the District of Columbia from 1983 through 2002, isolates from 21 cases of HUS were identified [43]. Serogroup O111 accounted for 48% of HUS cases, but missing data limited the power of the analysis. In Germany, Austria, and Australia, O111 strains were most frequently associated with HUS [33, 47]. These and other data suggest that infection with serotypes O26, O103, and O145 may also be relatively more likely to precipitate HUS (table 1). Although it is not clear whether the pathogenesis of HUS following infection with non-o157 strains exactly mirrors that following infection due to O157 strains, non-o157 strains harboring the gene encoding stx2 (see Disease Pathogenesis: Complex and Evolving) are those most highly associated with onset of HUS, a finding analogous to that for O157 strains. This similarity suggests that some infections due to non-o157 strains trigger the same abnormalities of the clotting cascade that precipitate HUS following infection due to O157 strains [48]. More data are needed to determine whether, for example, peripheral leukocytosis signals a greater probability of HUS, as has been noted in O157-associated disease [49, 50]. DISEASE PATHOGENESIS: COMPLEX AND EVOLVING The ability to express one or more subtypes of Shiga toxin defines membership in the STEC family. Production of Shiga toxin is central to the pathogenesis of both O157- and non- O157 associated human disease. Shiga toxin 1 (Stx1) and Shiga toxin 2 (Stx2) are the main members of this toxin family. In contrast to STEC that produce Stx1 only, STEC that produce Stx2 alone or both Stx1 and Stx2 are more likely to be associated with HUS [51]. Of note, intravenous administration of purified Stx2 alone in the baboon model of HUS is sufficient to cause the hallmark pathological features of HUS, at doses at which Stx1 alone will not. Like cholera toxin and pertussis toxin, the Shiga toxins are members of the AB5 toxin family, consisting of 1 enzymatically active A subunit non-covalently associated with a binding pentamer of identical B subunits. The B subunit binds to neutral glycolipids on host cells and mediates cellular uptake of Shiga toxin with trafficking to the endoplasmic reticulum. In the endoplasmic reticulum, the A subunit is translocated across the endoplasmic reticulum membrane to the cytoplasm, thereby gaining access to cytoplasmic ribosomes. The A subunit recognizes and binds to the 28S RNA portion of the 60S ribosomal subunit and cleaves off a specific adenine, preventing aminoacyl t-rna binding. This results in inhibition of protein synthesis and can trigger host cell signal transduction mechanisms that may favor expression of proinflammatory cytokines. Therefore, Shiga toxins likely have 2 mechanisms by which they cause local and systemic disease: direct (inhibition of protein synthesis) and indirect (proinflammatory cytokine expression). Because STEC are noninvasive, Shiga toxins must be absorbed across the intestinal epithelium to gain access to the systemic circulation. It is thought that Shiga toxins, once absorbed, act on a variety of host cell subtypes, including endothelial cells in target organs, such as the kidney and the brain, to cause pathological hallmarks of systemic disease: deposition of fibrin-platelet thrombi in affected tissues. Recent data suggest that there may be additional mediators of cell injury and death, including the secreted zinc metalloprotease of C1 esterase inhibitor [52], subtilase cytotoxin (SubAB), and cytolethal distending toxin [53]. Thus, the degree to which different strains of STEC may colonize the gastrointestinal tract, cause disruption of the gastrointestinal barrier, and produce Shiga toxin locally may vary (thereby influencing the likelihood of developing HUS) according to strain-dependent encoded and expressed virulence genes. Virulence characteristics vary among non-o157 strains and likely influence their pathogenicity. STEC strains may carry the locus of enterocyte effacement (LEE), a 43-kb pathogenicity island containing a cluster of genes whose cooperative protein products contribute to STEC pathogenesis. The presence of LEE 1590 CID 2006:43 (15 December) EMERGING INFECTIONS

in non-o157 strains correlates with virulence, because strains containing LEE are more frequently associated with HUS and outbreaks than are LEE-negative strains. Consistent with the increased association between STEC carrying LEE and human disease, serogroups O26, O103, O111, O121, and O145 (all associated with outbreaks of severe disease, including HUS [54]) are predominantly LEE-positive [38, 54, 55] and cdtnegative strains [56]. These and other data suggest that infection with certain serogroups may be more likely to precipitate HUS (table 3). LEE contains genes whose expression permits formation of the attaching and effacing lesions characteristic of STEC intestinal colonization. LEE also contains genes that encode a type III secretion system that delivers the proteins mediating the attaching and effacing lesions and other virulence proteins to enterocytes. Attaching and effacing lesions form when LEE-encoded intimin (eae), an outer membrane protein homologue, binds to the translocated intimin receptor (tir) that inserts into the enterocyte membrane a unique example of a pathogen providing to the host the receptor required for its colonization. In a Finnish study of all human non-o157 strains collected over a decade from symptomatic persons, 70% carried the eae gene [29], 55% carried stx2, and 75% carried E-hly. E- hly encodes enterohemolysin, a pore-forming cytolysin that may be associated with the presence of eae in non-o157 versus O157 strains [29]. The precise contribution of enterohemolysin to STEC virulence remains unknown. However, none of these human isolates lacked all 3 of these virulence genes. In other work, the adherence of STEC to enterocytes (dependent, in part, on the expression of eae) correlated directly with pathogenicity when strains from sporadic and outbreak-associated HUS were compared with isolates from food and from species other than humans [29, 74]. Genes encoding potential virulence factors are found more frequently in non-o157 STEC serogroups isolated from patients with bloody and/or HUS than they are from patients with non-o157 STEC associated nonbloody. The observations that O157 and non-o157 LEE-negative STEC can lead to significant disease in humans led investigators to look for other virulence traits, resulting in the discovery of new pathogenicity islands. Notably, the high-pathogenicity island is unique to certain non-o157 serogroups, such as O26, one of the most frequently reported serogroups in Europe and the United States [75]. The high-pathogenicity island encodes genes required for synthesis of an iron-recruitment system, called yersiniabactin, that is also found in pathogenic Yersinia species and other enterobacteria. Sequencing of the O157 genome identified 8 additional pathogenicity-associated islands that may contribute to strain virulence [76]. Certain non-o157 strains known to cause severe disease in humans carry 1 of these recently recognized pathogenicity islands, but further work is needed to understand the relationship of these islands to both O157- and non-o157 related human disease [54]. Additional investigation into the determinants of virulence of non-o157 STEC will likely yield exciting results, as illustrated by the recent discovery of a novel AB5 toxin, SubAB, in an Australian STEC O113:H21 strain associated with a small outbreak of HUS [77]. Notably, this strain does not possess the eae gene. Intraperitoneal injection of purified SubAB is fatal in a murine model, causing microvascular thrombosis and the subsequent necrosis within multiple organs [77]. SubAB genes have been found in fecal extracts from 10% of Australian STEC cases and are typically found in STEC that produce only Stx2 [78]. The worldwide distribution of SubAB in both STEC and other E. coli is unknown, and its contribution to strain virulence and disease requires further study. SCREENING AND ISOLATION OF NON-O157: A DIFFICULT SEARCH Stool samples for cultures must be obtained early in the disease course to reliably isolate O157, and presumably the same is true for non-o157 strains. O157 strains are easier to detect on stool culture, because most strains fail to metabolize sorbitol. This leads to colorless, opaque-appearing colonies on sorbitol- MacConkey agar, rather than the pink colonies (caused by fermentation of sorbitol) typical of most non-o157 STEC, as well as commensal and pathogenic -inducing classes of E. coli (e.g., enteropathogenic E. coli and enterotoxigenic E. coli, among others). Thus, given the serogroup diversity of non-o157 STEC, the only practical rapid approach, at present, to identify non- O157 STEC related disease is direct or indirect detection of Shiga toxin production by these strains. One of 2 EIAs approved by the US Food and Drug Administration to detect Shiga toxin may be performed either directly on stool samples or on supernatants of an overnight enrichment culture of stool [79]. Sensitivity of EIA is enhanced by adding mitomycin C or polymyxin B to stool cultures to stimulate production or liberation of bacterial cell associated Shiga toxin. Alternatively, a tissue culture cytotoxicity and neutralization assay can detect Shiga toxin, but this reference standard approach is too laborious and costly to be widely adopted for diagnosis. PCR-based methods to detect Shiga toxin genes in l stools may emerge as a rapid identification method in diagnostic laboratories capable of reliably performing these assays. One limitation of these approaches is that the non- O157 strains are not recovered and characterized. Ultimately, culture remains important for tracking non-o157 disease epidemiology and the development of public health strategies for containment. Is rapid diagnosis of infection due to non-o157 STEC clinically necessary? Given the severe disease associated with some non-o157 STEC serogroups, rapid diagnostic measures con- EMERGING INFECTIONS CID 2006:43 (15 December) 1591

Table 3. Overview of reported non-o157 Shiga toxin producing Escherichia coli (STEC) serogroups and associated clinical syndromes. Table 3. (Continued.) STEC serogroup Nonbloody HUS Bloody Reference(s) STEC serogroup Nonbloody HUS Bloody Reference(s) O1 Yes Yes Yes [14, 44] O2 No Yes a Yes [30, 37, 44] O5 No Yes Yes [21, 26, 44] O8 Yes Yes Yes [16, 32, 37] O9 Yes No No [14] O15 No No Yes [26, 29, 37] O18 No Yes Yes [21, 57] O22 Yes Yes No [32, 41, 58] O23 No Yes Yes [16, 41] O25 No No Yes [37] O26 Yes Yes Yes [14, 33, 34, 36, 37, 41, 59 62] O28 No Yes Yes [15, 17, 43] O45 Yes No Yes [6, 41] O46 Yes Yes No [32, 34] O48 No Yes No [34] O50 No No Yes [44] O55 Yes Yes Yes [16, 20, 26, 41, 43, 63] O58 No No Yes [37] O68 No No Yes [64] O73 Yes Yes Yes [6, 41, 63] O75 b No Yes No [41] O76 Yes No Yes [1, 32, 58] O91 Yes Yes No [1, 26, 28, 32, 34, 37, 41] O92 No Yes No [16] O98 No Yes No [34] O100 b No No Yes [41] O101 a No Yes No [30] O103 Yes Yes Yes [1, 11, 19, 37, 39, 62] O104 Yes Yes Yes [34, 41, 65] O107 a No Yes No [30] O111 Yes Yes Yes [1, 14, 15, 18 20, 26, 33, 36, 37, 41, 43, 62, 63, 66 71] O113 Yes Yes Yes [1, 33, 37, 54, 71] O115 Yes Yes No [38] O117 No No Yes [26] O118 Yes Yes Yes [1, 16, 32] O119 Yes Yes Yes [14 16, 41, 43, 63, 70] O121 Yes Yes Yes [10, 16, 26, 32, 37, 63, 72] O126 b Yes No No [41] O128 Yes Yes Yes [1, 20, 39, 41, 43, 63, 72] O130 No Yes No [33] O132 b Yes No No [41] O133 Yes No No [14] O145 Yes Yes Yes [1, 10, 19, 32, 37, 39, 63] O146 Yes Yes Yes [1, 26, 34] O165 No Yes Yes [6, 15, 16, 34, 39, 43] O166 Yes No No [41] O171 No No Yes [37] O174 Yes Yes a Yes [26, 30, 37] (continued) O177 No Yes Yes [6, 73] O179 Yes No Yes [73] O180 Yes No No [73] O181 Yes No Yes [6, 73] OR Yes Yes Yes [41, 6, 32, 33, 16, 30, 39] ONT Yes Yes Yes [1, 6, 14, 16, 32, 33, 37 39, 41] NOTE. Data are derived from case series, outbreak reports, and case reports. Strains reported without clinical information are not included. HUS, hemolytic-uremic syndrome; ONT, O nontypeable; OR, O rough. a Single strain for a patient presenting with either HUS or thrombotic thrombocytopenic purpura; prior clinical syndrome was not specified [30]. b Single strains reported for serogroup. firmed by culture are optimal. Although testing all patients with l syndromes for non-o157 STEC infection is ideal, costs may well be prohibitive in regions in which incidence is low. For this reason, development and investigations of rapid, cost-effective diagnostic methods are imperative. The leading life-threatening sequela of STEC infection is development of HUS. Because the critical pathophysiological changes of HUS are occurring by the time that the patient presents with the l prodrome, only early interventions requiring rapid diagnostics to interrupt coagulation cascade dysregulation will reduce morbidity and mortality caused by HUS. Further, bloody is a serious illness among both children and adults, compelling clinicians to consider early, empirical antibiotic therapy. However, current recommendations and available data (although limited in scope and only formally studied for O157-related infections in children) suggest that antibiotics should be withheld if STEC infection is suspected, given concerns that antibiotics may trigger release of Stx and progression to HUS, resulting in worse clinical outcomes. These considerations support an argument for early diagnosis of STEC infection. The case for rapid specific diagnosis of STEC infection is further supported by a recent prospective study of 29 children with O157-associated bloody or HUS, which suggests that early isotonic volume expansion occurring before culture results are available may attenuate renal injury and failure [80]. The value and potential complications (e.g., volume overload) of aggressive hydration in adults and the benefit in non- O157 STEC infections are unknown and are likely to remain so, indicating that clinicians may well have to take their management cues from these limited data. Thus, rapid diagnosis of all STEC infections should assist clinical decision-making regarding fluid and antibiotic administration. Because the clinical manifestations of STEC infection may mimic appendicitis, intussusception, inflammatory bowel disease, and Clostridium 1592 CID 2006:43 (15 December) EMERGING INFECTIONS

difficile infection, rapid diagnosis may avert both invasive endoscopy procedures, which are costly and carry potential morbidity, and emergent surgical procedures. Finally, rapid diagnosis may serve to limit outbreaks. Clinician-initiated reports of STEC infection to public health authorities will allow timely outbreak investigations that may permit detection of additional STEC cases, limit spread from persistently contaminated common sources, and/or prevent secondary spread within families. CONCLUSIONS AND RECOMMENDATIONS Non-O157 STEC infections may induce a range of illness, from mild gastroenteritis to critical illnesses, including hemorrhagic colitis, HUS, and death, either as sporadic cases or in outbreaks, reported in the United States and internationally. A substantial proportion of HUS is caused by infection due to non-o157 strains worldwide, but the true incidence and burden of illness caused by these E. coli serotypes remain unknown. In 2000, only 68% of clinical laboratories in the United States tested routinely for infections due to O157 STEC, with fewer laboratories testing for non-o157 STEC infections [81]. The search for STEC infection should begin, at a minimum, in patients presenting with bloody and/or HUS. Ideally, screening for STEC on all stool specimens submitted because of l illness should be considered, but the cost-effectiveness of this approach, given the geographic variability of both O157 and non-o157 STEC infections, requires assessment. If specific and effective therapies that prevent HUS are developed, the need to diagnose STEC infection will be compelling. Rapid, sensitive, accurate, and inexpensive techniques to enable routine testing for Shiga toxin and non-o157 strains should be developed and incorporated into clinical laboratory protocols. Only through active case investigation will we further understand the distribution and human cost of non-o157 associated disease. Acknowledgments We appreciate the helpful input received from Drs. Karen Carroll (Johns Hopkins University School of Medicine, Baltimore, MD), James Kaper (University of Maryland School of Medicine, Baltimore, MD), and David Acheson (US Food and Drug Administration, Rockville, MD). Financial support. National Institutes of Health (T32 AI07291 to K.E.J.; AI059509 to C.M.T.; and DK/AI59655 and DK45496 to C.L.S.), and the Crohn s and Colitis Foundation of America (to C.L.S.). Potential conflicts of interest. C.M.T. has received financial support in the form of patent royalties from Meridian Diagnostics. K.E.J. and C.L.S.: no conflicts. References 1. Beutin L, Zimmermann S, Gleier K. Human infections with Shiga toxin producing Escherichia coli other than serogroup O157 in Germany. Emerg Infect Dis 1998; 4:635 9. 2. Goldwater PN, Bettelheim KA. Role of non-o157:h7 Escherichia coli in hemolytic uremic syndrome. Clin Infect Dis 2002; 35:346 7. 3. Acheson DWK, Keusch GT. Which Shiga toxin producing types of E. coli are important? ASM News 1996; 62:302 6. 4. Mead PS, Slutsker L, Dietz V, McCaig LF, Bresee JS, Shapiro C. Foodrelated illness and death in the United States. Emerg Infect Dis 1999;5: 607 25. 5. Andreoli SP, Trachtman H, Acheson DWK, Siegler RL, Obrig TG. Hemolytic uremic syndrome: epidemiology, pathophysiology, and therapy. Pediatr Nephrol 2002; 17:293 8. 6. Jelacic JK, Damrow T, Chen GS, et al. Shiga toxin producing Escherichia coli in Montana: bacterial genotypes and clinical profiles. J Infect Dis 2003; 188:719 29. 7. Banatvala N, Griffin PM, Greene KD, et al. The United States National Prospective Hemolytic Uremic Syndrome Study: microbiologic, serologic, clinical, and epidemiologic findings. J Infect Dis 2001; 183: 1063 70. 8. Fey PD, Wickert RS, Rupp ME, Safranek TJ, Hinrichs SH. Prevalence of non-o157:h7 Shiga toxin producing Escherichia coli in l stool samples from Nebraska. Emerg Infect Dis 2000; 6:530 3. 9. Pai CH, Ahmed N, Lior H, Johnson WM, Sims HV, Woods DE. Epidemiology of sporadic due to verocytotoxin-producing Escherichia coli: a 2-year prospective study. J Infect Dis 1988; 157:1054 7. 10. Karmali MA, Petric M, Lim C, Fleming PC, Arbus GS, Lior H. The association between idiopathic hemolytic uremic syndrome and infection by verotoxin-producing Escherichia coli. J Infect Dis 1985; 151: 775 82. 11. Ramotar K, Henderson E, Szumski R, Louie TJ. Impact of free verotoxin testing on epidemiology of caused by verotoxin-producing Escherichia coli. J Clin Microbiol 1995; 33:1114 20. 12. Ritchie M, Partington S, Jessop J, Kelly MT. Comparison of a direct fecal Shiga-like toxin assay and sorbitol-macconkey agar culture for laboratory diagnosis of enterohemorrhagic Escherichia coli infection. J Clin Microbiol 1992; 30:461 4. 13. Kleanthous H, Smith HR, Scotland SM, et al. Haemolytic uraemic syndromes in the British Isles, 1985 8: association with verocytotoxin producing Escherichia coli. Arch Dis Child 1990; 84:722 7. 14. Huppertz HI, Busch D, Schmidt H, Aleksic S, Karch H. Diarrhea in young children associated with Escherichia coli non-o157 organisms that produce Shiga-like toxin. J Pediatr 1996; 128:341 6. 15. Gerber A, Karch H, Allerberger F, Verweyen HM, Zimmerhackl LB. Clinical course and the role of Shiga toxin producing Escherichia coli infection in the hemolytic-uremic syndrome in pediatric patients, 1997 2000, in Germany and Austria: a prospective study. J Infect Dis 2002; 186:493 500. 16. Friedrich AW BM, Zhang W, Pulz M, Kuczius T, Ammon A, Karch H. Escherichia coli harboring Shiga toxin 2 gene variants: frequency and association with clinical symptoms. J Infect Dis 2002; 185:74 84. 17. Blanco JE, Blanco M, Alonso MP, et al. Serotypes, virulence genes, and intimin types of Shiga toxin (verotoxin) producing Escherichia coli isolates from human patients: prevalence in Lugo, Spain, from 1992 through 1999. J Clin Microbiol 2004; 42:311 9. 18. Caprioli A, Luzzi I, Rosmini F, et al. Communitywide outbreak of hemolytic-uremic syndrome associated with non-o157 verocytotoxinproducing Escherichia coli. J Infect Dis 1994; 169:208 11. 19. Tozzi AE, Caprioli A, Minelli F, et al. Shiga toxin producing Escherichia coli infections associated with hemolytic uremic syndrome, Italy, 1988 2000. Emerg Infect Dis 2003; 9:106 8. 20. Bielaszewska M, Janda JM, Blahova K, Feber J, Potuznik V, Souckova A. Verocytotoxin-producing Escherichia coli in children with hemolytic uremic syndrome in the Czech Republic. Clin Nephrol 1996; 46:42 4. 21. Sramkova L, Bielaszewska M, Janda J, Blahova K, Hausner O. Vero cytotoxin producing strains of Escherichia coli in children with haemolytic uraemic syndrome and diarrhoea in Czechoslovakia. Infection 1990; 18:204 9. 22. Pierard D, Stevens D, Moriau L, Lior H, Lauwers S. Three years PCR screening for VTEC in human stools in Brussels. In: Karmali MA, Goglio AG, eds. Recent advances in verocytotoxin-producing Escherichia coli infections. Amsterdam: Elsevier Science, 1994:33 6. 23. Bonnet R, Souweine B, Gauthier G, et al. Non-O157:H7 Stx2-producing Escherichia coli strains associated with sporadic cases of hemolytic-uremic syndrome in adults. J Clin Microbiol 1998; 36:1777 80. EMERGING INFECTIONS CID 2006:43 (15 December) 1593

24. Decludt B, Bouvet P, Mariani-Kurkdjian P, et al. Haemolytic uraemic syndrome and Shiga toxin producing Escherichia coli infection in children in France. The Societe de Nephrologie Pediatrique. Epidemiol Infect 2000; 124:215 20. 25. Pradel N, Boukhors K, Bertin Y, Forestier C, Martin C, Livrelli V. Heterogeneity of Shiga toxin producing Escherichia coli strains isolated from hemolytic-uremic syndrome patients, cattle, and food samples in central France. Appl Environ Microbiol 2001; 67:2460 8. 26. Ethelberg S, Olsen KE, Scheutz F, et al. Virulence factors for hemolytic uremic syndrome, Denmark. Emerg Infect Dis 2004; 10:842 7. 27. Nielsen EM, Scheutz F, Torpdahl M. Continuous surveillance of Shiga toxin producing Escherichia coli infections by pulsed-field gel electrophoresis shows that most infections are sporadic. Foodborne Pathog Dis 2006; 3:81 7. 28. Keskimaki M, Saari M, Heiskanen T, Siitonen A. Shiga toxin producing Escherichia coli in Finland from 1990 through 1997: prevalence and characteristics of isolates. J Clin Microbiol 1998; 36:3641 6. 29. Eklund M, Scheutz F, Siitonen A. Clinical isolates of non-o157 Shiga toxin producing Escherichia coli: serotypes, virulence characteristics, and molecular profiles of strains of the same serotype. J Clin Microbiol 2001; 39:2829 34. 30. Eklund M. Enterohemorrhagic Escherichia coli (EHEC) findings from humans in Finland [academic dissertation]. Helsinki: National Public Health Institute, 2005. 31. Eklund M, Nuorti JP, Ruutu P, Siitonen A. Shigatoxigenic Escherichia coli (STEC) infections in Finland during 1998 2002: a populationbased surveillance study. Epidemiol Infect 2005; 133:845 52. 32. Welinder-Olsson C, Badenfors M, Cheasty T, Kjellin E, Kaijser B. Genetic profiling of enterohemorrhagic Escherichia coli strains in relation to clonality and clinical signs of infection. J Clin Microbiol 2002; 40: 959 64. 33. Elliott EJ, Robins-Browne RM, O Loughlin EV, et al. Nationwide study of haemolytic uraemic syndrome: clinical, microbiological, and epidemiological features. Arch Dis Child 2001; 85:125 31. 34. Goldwater PN, Bettelheim KA. The role of enterohaemorrhagic E. coli serotypes other than O157:H7 as causes of disease. In: Karmali MA, Goglio AG, eds. Recent advances in verocytotoxin-producing Escherichia coli infections. Amsterdam: Elsevier Science, 1994:57 60. 35. Brooks HJ, Bettelheim KA, Todd B, Holdaway MD. Non-O157 Vero cytotoxin producing Escherichia coli: aetiological agents of diarrhoea in children in Dunedin, New Zealand. Comp Immunol Microbiol Infect Dis 1997; 20:163 70. 36. Cordovez A, Prado V, Maggi L, et al. Enterohemorrhagic Escherichia coli associated with hemolytic-uremic syndrome in Chilean children. J Clin Microbiol 1992; 30:2153 7. 37. Rivas M, Miliwebsky E, Chinen I, et al. Characterization and epidemiologic subtyping of Shiga toxin producing Escherichia coli strains isolated from hemolytic uremic syndrome and cases in Argentina. Foodborne Pathog Dis 2006; 3:88 96. 38. Kudoh Y, Kai A, Obata H, et al. Epidemiological surveys on verocytotoxin-producing Escherichia coli infections in Japan. In: Karmali MA, Goglio AG, eds. Recent advances in verocytotoxin-producing Escherichia coli infections. Amsterdam: Elsevier Science, 1994:53 6. 39. Uchida H, Kanegane H, Yoshiya K, Kitamura K, Ihara T, Kamiya H. Four cases of hemolytic uremic syndrome (HUS) associated with serotype O165 verotoxin producing Escherichia coli [in Japanese]. Kansenshogaku Zasshi 1995; 69:678 83. 40. Bitzan M, Karch H, Maas MG, et al. Clinical and genetic aspects of Shiga-like toxin production in traditional enteropathogenic Escherichia coli. Zentralbl Bakteriol 1991; 274:496 506. 41. Bockemuhl J, Aleksic S, Karch H. Serological and biochemical properties of Shiga-like toxin (verocytotoxin) producing strains of Escherichia coli, other than O-group 157, from patients in Germany. Zentralbl Bakteriol 1992; 276:189 95. 42. Lopez EL, Diaz M, Grinstein S, et al. Hemolytic uremic syndrome and in Argentine children: the role of Shiga-like toxins. J Infect Dis 1989; 160:469 75. 43. Brooks JT, Sowers EG, Wells JG, et al. Non-O157 Shiga toxin producing Escherichia coli infections in the United States, 1983 2002. J Infect Dis 2005; 192:1422 9. 44. Bielaszewska M, Janda JM, Blahova K, Sramkova L, Havlik J, Potuznik V. Verocytotoxin-producing Escherichia coli in children with hemolytic uremic syndrome and in the Czech Republic. In: Karmali MA, Goglio AG, eds. Recent advances in verocytotoxin-producing Escherichia coli infections. Amsterdam: Elsevier Science, 1994:37 40. 45. Werber D, Fruth A, Heissenhuber A, et al. Shiga toxin producing Escherichia coli O157 more frequently cause bloody than do non-o157 strains. J Infect Dis 2004; 189:1335 6. 46. Garg AX, Suri RS, Barrowman N, et al. Long-term renal prognosis of -associated hemolytic uremic syndrome. JAMA 2003; 290: 1360 70. 47. Gerber A, Karch H, Allerberger F, Verweyen HM, Zimmerhackl LB. Clinical course and the role of Shiga toxin producing Escherichia coli infection in the hemolytic-uremic syndrome in pediatric patients, 1997 2000, in Germany and Austria: a prospective study. J Infect Dis 2002; 186:493 500. 48. Chandler WL, Jelacic S, Boster DR, et al. Prothrombotic coagulation abnormalities preceding the hemolytic-uremic syndrome. N Engl J Med 2002; 346:23 32. 49. Dundas S, Todd WTA, Stewart AI, Murdoch PS, Chaudhuri AKR, Hutchinson SJ. The central Scotland Escherichia coli O157:H7 outbreak: risk factors for the hemolytic uremic syndrome and death among hospitalized patients. Clin Infect Dis 2001; 33:923 31. 50. Wong CS, Jelacic S, Habeeb RL, Watkins SL, Tarr PI. The risk of the hemolytic-uremic syndrome after antibiotic treatment of Escherichia coli O157:H7 infections. N Engl J Med 2000; 342:1930 6. 51. Ostroff SM, Tarr PI, Neill MA, Lewis JH, Hargrett-Bean N, Kobayashi JM. Toxin genotypes and plasmid profiles as determinants of systemic sequelae in Escherichia coli O157:H7 infections. J Infect Dis 1989; 160: 994 8. 52. Lathem WW, Grys TE, Witowski SE, et al. StcE, a metalloprotease secreted by Escherichia coli O157:H7, specifically cleaves C1 esterase inhibitor. Mol Microbiol 2002; 45:277 88. 53. Karch H, Friedrich AW, Gerber A, Zimmerhackl LB, Schmidt MA, Bielaszewska M. New aspects in the pathogenesis of enteropathic hemolytic uremic syndrome. Semin Thromb Hemost 2006; 32:105 12. 54. Karmali MA, Mascarenhas M, Shen S, et al. Association of genomic O island 122 of Escherichia coli EDL 933 with verocytotoxin-producing Escherichia coli seropathotypes that are linked to epidemic or serious disease. J Clin Microbiol 2003; 41:4930 40. 55. Boerlin P, McEwen SA, Boerlin-Petzold F, Wilson JB, Johnson RP, Gyles CL. Associations between virulence factors of Shiga toxin producing Escherichia coli and disease in humans. J Clin Microbiol 1999; 37: 497 503. 56. Bielaszewska M, Fell M, Greune L, et al. Characterization of cytolethal distending toxin genes and expression in Shiga toxin producing Escherichia coli strains of non-o157 serogroups. Infect Immun 2004; 72: 1812 6. 57. Bielaszewska M, Sramkova L, Janda J, Blahova K, Ambrozova H. Verotoxigenic (enterohaemorrhagic) Escherichia coli in infants and toddlers in Czechoslovakia. Infection 1990; 18:352 6. 58. Welinder-Olsson C, Eriksson E, Kaijser B. Virulence genes in verocytotoxigenic Escherichia coli strains isolated from humans and cattle. APMIS 2005; 113:577 85. 59. Tozzi AE, Caprioli A, Minelli F, et al. Shiga toxin producing Escherichia coli infections associated with hemolytic uremic syndrome, Italy, 1988 2000. Emerg Infect Dis 2003; 9:106 8. 60. Werber D, Fruth A, Liesegang A, et al. A multistate outbreak of Shiga toxin producing Escherichia coli O26:H11 infections in Germany, detected by molecular subtyping surveillance. J Infect Dis 2002; 186: 419 22. 61. Allerberger F, Friedrich AW, Grif K, et al. Hemolytic-uremic syndrome associated with enterohemorrhagic Escherichia coli O26:H infection and 1594 CID 2006:43 (15 December) EMERGING INFECTIONS

consumption of unpasteurized cow s milk. Int J Infect Dis 2003;7: 42 5. 62. Schmidt H, Geitz C, Tarr PI, Frosch M, Karch H. Non-O157:H7 pathogenic Shiga toxin producing Escherichia coli: phenotypic and genetic profiling of virulence traits and evidence for clonality. J Infect Dis 1999; 179:115 23. 63. Bitzan M, Karch H, Maas MG, et al. Clinical and genetic aspects of Shiga-like toxin production in traditional enteropathogenic Escherichia coli. Zentralbl Bakteriol 1991; 274:496 506. 64. Bokete TN, O Callahan CM, Clausen CR, et al. Shiga-like toxin producing Escherichia coli in Seattle children: a prospective study. Gastroenterology 1993; 105:1724 31. 65. Outbreak of acute gastroenteritis attributable to Escherichia coli serotype O104:H21 Helena, Montana, 1994. MMWR CDC Surveill Summ 1994; 44:501 3. 66. Community outbreak of hemolytic uremic syndrome attributable to Escherichia coli O111:NM South Australia, 1995. MMWR CDC Surveill Summ 1995; 44:550 1. 67. Bitzan M, Karch H, Altrogge H, Strehlau J, Blaker F. Hemolytic-uremic syndrome associated with a variant Shiga-like cytotoxin of Escherichia coli O111. Pediatr Infect Dis J 1988; 7:128 32. 68. Gunzburg S, Gracey M, Forbes D, Hewitt I, Bettelheim K. Haemolyticuraemic syndrome and verocytotoxigenic Escherichia coli. Med J Aust 1988; 149:54 5. 69. Guth BEC, Ramos SRTS, Cerqueira AMF, Andrade JRC, Gomes TAT. Phenotypic and genotypic characteristics of Shiga toxin producing Escherichia coli strains isolated from children in São Paulo, Brazil. Mem Inst Oswaldo Cruz 2002; 97:1085 9. 70. Brooks JT, Bergmire-Sweat D, Kennedy M, et al. Outbreak of Shiga toxin producing Escherichia coli O111:H8 infections among attendees of a high school cheerleading camp. Clin Infect Dis 2004; 38:190 8. 71. Banatvala N, Debeukelaer MM, Griffin PM, et al. Shiga-like toxin producing Escherichia coli O111 and associated hemolyticuremic syndrome: a family outbreak. Pediatr Infect Dis J 1996; 15: 1008 11. 72. McCarthy TA, Barrett NL, Hadler JL, et al. Hemolytic-uremic syndrome and Escherichia coli O121 at a lake in Connecticut, 1999. Pediatrics 2001; 108:1 7. 73. Scheutz F, Cheasty T, Woodward D, Smith HR. Designation of O174 and O175 to temporary O groups OX3 and OX7, and six new Escherichia coli O groups that include verocytotoxin-producing E. coli (VTEC): O176, O177, O178, O179, O180, and O181. APMIS 2004; 112:569 84. 74. Paton AW, Voss E, Manning PA, Paton JC. Shiga toxin producing Escherichia coli isolates from cases of human disease show enhanced adherence to intestinal epithelial (Henle 407) cells. Infect Immun 1997; 65:3799 805. 75. Karch H, Schubert S, Zhang D, et al. A genomic island, termed highpathogenicity island, is present in certain non-o157 Shiga toxin producing Escherichia coli clonal lineages. Infection and Immunity 1999; 67:5994 6001. 76. Perna NT, Plunkett G III, Burland V, et al. Genome sequence of enterohaemorrhagic Escherichia coli O157:H7. Nature 2001; 409:529 33. 77. Paton AW, Srimanote P, Talbot UM, Wang H, Paton JC. A new family of potent AB5 cytotoxins produced by Shiga toxigenic Escherichia coli. J Exp Med 2004; 200:35 46. 78. Paton AW, Paton JC. Multiplex PCR for direct detection of Shiga toxigenic Escherichia coli strains producing the novel subtilase cytotoxin. J Clin Microbiol 2005; 43:2944 7. 79. Carroll KC. Immunoassay detection of Shiga toxin producing Escherichia coli. In: Isenberg H, ed. Clinical microbiology procedures handbook. Washington, DC: American Society for Microbiology Press (in press). 80. Ake JA, Jelacic S, Clark C, et al. Relative nephroprotection during Escherichia coli O157:H7 infections: association with intravenous volume expansion. Pediatrics 2005; 115:673 80. 81. Voetsch AC, Angulo FJ, Rabatsky-Ehr T, et al. Laboratory practices for stool-specimen culture for bacterial pathogens, including Escherichia coli O157:H7, in the FoodNet sites, 1995 2000. Clin Infect Dis 2004; 38(Suppl 3):S190 7. EMERGING INFECTIONS CID 2006:43 (15 December) 1595