Extraintestinal Pathogenic Escherichia coli Carrying the Shiga Toxin Gene stx₂

LETTER

The 2011 outbreak of infections with Escherichia coli with characteristics of both enteroaggregative E. coli (EAEC) and Shiga toxin-producing E. coli (STEC) caused a paradigm shift with regard to the human pathogenicity of STEC (1). Diarrheagenic E. coli (DEC) usually does not cause extraintestinal diseases such as urinary tract infections and bacteremia (2).

The DNA of 193 E. coli isolates from adult bacteremic patients described previously (3) was analyzed for the presence of stx₁ and stx₂ (4). All of the strains were stx₁ negative, but nine (4.7%) were stx₂ positive. According to the Scheutz subtyping protocol (5), seven belonged to subtype stx_2c and two belonged to stx_2d. Table 1 shows the associations between the clinical presentations of the patients and their stx₂ statuses.

Table 1.

stx₂ statuses of E. coli and clinical characteristics of 193 bacteremia patients

Characteristic	No. (%) of patients			P value
	All	184 stx negative	9 stx positive
Age of ≥65 yr	133 (68.9)	126 (68.5)	7 (77.8)	0.56
Male gender	71 (36.8)	66 (35.9)	5 (55.6)	0.23
Symptoms of gastroenteritis	74 (38.1)	74 (40.4)	0	0.03
Clinical presentationa within 1 day of admission with:
Acute renal failure	10 (5.2)	9 (4.7)	1 (12.5)	0.35
Thrombocytopeniab	18 (9.5)	18 (9.9)	0	0.25
One or more failing organs	86 (45.7)	83 (46.1)	3 (37.5)	0.63
Death in hospital within 14 days after admission	14 (7.3)	13 (7.1)	1 (12.5)	0.65

^aData were available on acute renal failure, thrombocytopenia, and one or more failing organs for 191, 190, and 188 patients, respectively.

^b<100,000 thrombocytes/μl.

Arbitrarily chosen, DNA from 2 of the stx₂-postive strains and 26 of the stx-negative strains was analyzed for additional DEC genes (4, 6) and extraintestinal pathogenic E. coli (ExPEC) genes by PCR assay as selected by B. A. Lindstedt et al. (unpublished data) (Table 2). One of the two stx₂-positive strains that were further analyzed for additional virulence markers carried both stx_2c and aggR. Seven of the 26 stx-negative strains carried other DEC genes; with or without concomitant ExPEC genes.

Table 2.

Molecular characterization of a selection of E. coli bacteremia strains

Study identification no.	O serogroup	DEC VGsa	ExPEC VGsb
			Adhesins	Iron acquisition	Cytotoxins	Other
11	O?			sitA, iroC		ibeA, iss, traT, tsh, kps1
52	O2		papC	iutA, sitA, iucD, iroC		iss, traT, tsh, etsA, kps1
81	O103c	stx2c	papC	iutA, sitA, iucD, iroC		kps1, tsh, etsA, iss, traT
100	O?	ehaA
282	O?			iutA, sitA, iucD	sat	iss, traT
325	O6		sfaS	sitA, iroC	cnf1	iss, traT, tsh, kps1
336	O?	ehaA		iutA, iucD		iss, etsA
387	O?			iutA, iucD, iroC		iss, traT, tsh, etsA
447	O2		papC	iutA, sitA, iucD, iroD		iss, traT, tsh, etsA, kps1
585	O2	cdt		iutA, iucD, sitA, iroC	cnf1, sat	iss, kps1, tsh
668	O2		papC	iutA, sitA, iroC, iucD		iss, traT, tsh, etsA
685	O15	stx2c, aggR		iutA, sitA, iucD	sat	iss, traT
687	O12			iutA, iucD, iroC		iss, traT, tsh, etsA
713	O6/O7			iutA, sitA, iucD, iroC		iss, sat, tsh
803	O?			iutA, sitA, iucD	sat	ibeA, iss
804	O6			sitA, iroC		ibeA, iss, traT, tsh, kps1
839	O103c		papC	iutA, sitA, iucD, iroC		iss, traT, tsh, etsA, kps1
859	O?	ehaA		sitA, iucD	sat	iss, traT
865	O?	ehaA		iutA, iucD, icoC		iss, traT, etsA
891	O2			sitA		iss, etsA, kps1
895	O4/O12			iutA, sitA, iucD, iroC	cnf1	iss, tsh, kps1
915	O75			sitA, iroC	cnf1	ibeA, iss, tsh, kps1
952	O?			sitA, iroC		ibeA, iss, tsh, kps1
972	O18	cdt	sfaS	iutA, sitA, iucD, iroC		ibeA, iss, traT, gimB, tsh, etsA, kps1
1010	O6			iutA, sitA, iucD, iroC	cnf1, sat	iss, traT,tsh
1095	O1		papC	sitA		iss, traT, tsh, kps1
1121	O103c	eaeB, ehaA
1127	O?			sitA, iroC		ibeA, iss, traT, tsh, kps1

^aVirulence genes (VGs) associated with DEC: ST1a, ST1b, ehxA, aggR, LT1, stx₁, stx₂, eaeB, ipaH, bfpB, saa, nleB, stcE, cdt, and subA.

^bVGs associated with ExPEC: adhesins (papC and sfaS), iron acquisition (iutA, sitA, iucD, and iroC), cytotoxins (cnf1, cnf2, cnf3, and sat), and others (kps1, ibeA, iss, traT, and tsh).

^cFlagellar antigen H2.

The stx₂-positive strains were recultured and tested for toxin production by ImmunoCard STAT! EHEC (Meridian Bioscience, Inc.). All nine stx₂-positive strains tested negative for toxin production. Therefore, another isolation of DNA from the viable strains was performed and new PCR assays detecting stx and aggR were run. Surprisingly, the toxin gene was lost by all nine but aggR was reproduced in the relevant strain, indicating no mix-up of strains or DNA. Loss of stx₂-carrying bacteriophages is not a novel phenomenon (7, 8), but to our knowledge, such a frequent loss has not been described previously.

Reports of bacteremia or sepsis in patients with hemolytic-uremic syndrome (HUS) caused by STEC do exist (2, 9–11). In our study, we found that almost 5% of the strains found in adult patients with bacteremia caused by E. coli carried stx₂ subtypes associated with HUS (12–14). Not having symptoms of gastroenteritis was associated with stx₂-positive status, but this finding seems rather implausible. Further characterization of two stx_2c-carrying isolates showed features of STEC and EAEC, as well as ExPEC, which, to our knowledge, is a novel finding.

If reproduced, these stx findings may have consequences for infection control. And if an association with clinical presentation is found, differential diagnoses of bacteremia with E. coli should include STEC colitis, as well as HUS. This is particularly relevant for elderly patients, who may have vague symptoms and comorbidity complicating the clinical picture. Furthermore, elderly, institutionalized patients have a unique susceptibility to STEC infection and its sequelae (15). Microbiological analyses of E. coli blood culture isolates to detect stx should therefore be encouraged, and infection control measures and contact tracing should be implemented when stx-carrying E. coli bacteremia is confirmed.