Secondary Logo

Journal Logo

Pathogenesis and Host Response

Escherichia coli Bacteremia in Children

Age and Portal of Entry Are the Main Predictors of Severity

Burdet, Charles MD, MPH*†; Clermont, Olivier PhD; Bonacorsi, Stéphane MD, PhD‡§; Laouénan, Cédric MD, MPH*†; Bingen, Edouard PharmD‡§; Aujard, Yannick MD‡¶; Mentré, France MD, PhD*†; Lefort, Agnès MD, PhD†‖; Denamur, Erick MD, PhD for the COLIBAFI Group

Author Information
The Pediatric Infectious Disease Journal: August 2014 - Volume 33 - Issue 8 - p 872-879
doi: 10.1097/INF.0000000000000309
  • Free


Severe sepsis in children remains a common cause of hospitalization.1 Bacteremia account for 25% of hospitalization for severe bacterial infections, and mortality rate is as high as 17%.1 Few temporal changes in the distribution of bacterial species in bloodstream infections have been reported.2 In industrialized countries, Streptococcus pneumoniae, Staphylococcus aureus and Escherichia coli figure among the 5 more frequently isolated organisms.3

E. coli is a commensal bacteria of the gut.4 Some strains can cause however intestinal or extraintestinal infections because of specific virulence factors.5 The structure of E. coli population is mainly clonal,6 with 7 principal phylogenetic groups (A, B1, B2, C, D, E and F) and numerous clonal complexes.4,7–10 Most clinical extraintestinal pathogenic E. coli belong to the B2 phylogroup and to a lesser extent to the D phylogenetic group.11 Some clones or clonal complexes have been linked to specific syndromes.12,13

In adults, mortality of E. coli bacteremia has been reported to be as high as 13% in a recent study performed by our group.14 We showed that host factors and the portal of entry are the main drivers of mortality during E. coli bacteremia. The only bacterial characteristic influencing the prognosis was the presence of virulence factor ireA, which was negatively associated with death. Phylogenetic belonging of the strains was not a significant predictor.

Studies of E. coli bloodstream infections in children, especially in urinary-source bacteremia, linked some virulence factors to the development of bacteremia.15–17 They include adhesins, iron uptake systems, protectins and toxins, with differences according to the portal of entry. A higher content in virulence factors was thus reported in strains causing urosepsis when compared with strains causing urinary tract infections without bacteremia.16 However, there is a lack of prospective studies taking into account both host and bacterial characteristics versus the clinical outcome of E. coli bacteremia in children.

To better understand the pathophysiology and severity of E. coli bacteremia in children, we analyzed the cohort of children ≤18 years of age of the COLIBAFI study.14 We compared bacterial characteristics (1) in bacteremia from urinary versus digestive origin, (2) in children ≤3 versus >3 months of age and (3) in community-acquired urinary-source bacteremia in children versus adults. Then we searched for clinical and bacterial risk factors associated with the severity of E. coli bacteremia in children.


Study Design and Setting

This study is part of the COLIBAFI study, a large prospective observational study conducted in 15 French hospitals between January 2005 and November 2007 in which 1051 adults and 84 children with E. coli bacteremia were enrolled. The methodology has been previously published.14E. coli bacteremia was defined as the presence of E. coli in ≥1 aseptically filled blood culture vial(s). The follow up ended at hospital discharge or at day 28 after the first positive blood culture. The study was approved by the institutional Ethics Committee (Hôpital Saint-Louis, Paris, France).

Clinical and Bacterial Characteristics

Clinical characteristics included age, gender, birth weight, preterm birth, immunodepression and origin of infection. Portal of entry, immunodepression and nosocomial bacteremia were defined as in Lefort et al.14 Severity of bacteremia was defined as the occurrence of death or transfer to ICU during follow up.

Phylogenetic group of strains was determined by a quadruplex polymerase chain reaction (PCR) method derived from the Clermont method,18 associated to allele-specific PCRs allowing the delineation of 7 main phylogroups (ie, A, B1, B2, C, D, E and F).19 Strains exhibiting the A1 genotype (chuA−, yjaA+, TspE4.C2−)18 correspond to ST complex (STc) 10 according to the Achtman multilocus sequence typing scheme.9 Among the D phylogroup strains, clonal group A (CGA) was identified as previously described.20 Ten main B2 phylogenetic subgroups (I to X) were detected by allele-specific PCRs.8,21 They correspond to STc 131, 73, 127, 141, 144, 12, 14, 452, 95 and 372, respectively.9 O-typing was performed by PCR.22,23

The presence of 20 extraintestinal virulence factors was tested by PCR,14,24 including adhesins/invasins (papC, papG II and III alleles, sfa/foc, iha, hra and ibeA), toxins (hlyC, cnf1, sat, clbA and clbQ), iron capture systems (fyuA, irp2, iroN, iucC, and ireA), protectins (neuC, chromosomal ompT, and traT) and a gene encoding an uropathogenic-specific protein, usp. The presence of 5 intestinal virulence genes was also tested by PCR23: afaD (diffusely adherent E. coli), eltB and estA (enterotoxigenic E. coli), eae (enteropathogenic E. coli and enterohemorrhagic E. coli) and aatA (enteroaggregative E. coli). The presence of 7 pathogenicity-associated islands (PAIs) was deduced from the presence of virulence factors25,26: PAI ICFT073 (papGII, hlyC and iucC positive), PAI IIJ96 (presence of at least 3 of the 4 following genes: papGIII, hlyC, cnf1 and hra), PAI III536 (sfa/foc and iroN positive), PAI IV536 (irp2 and fyuA positive), PAIgimA (ibeA positive), PAIUSP (usp positive) and PAIpks (clbA and clbQ positive). For each strain, virulence and PAIs scores were computed as the number of virulence factors and PAIs present.

Antimicrobial susceptibility was assessed using the disk diffusion method, as recommended by the Comité de l’Antibiogramme de la Société Française de Microbiologie ( Intermediate susceptibility was regarded as resistance. Resistance to cefotaxime and/or ceftazidime defined resistance to third-generation cephalosporin. Resistance to amoxicillin, ofloxacin and cotrimoxazole defined multidrug resistance. A resistance score was computed based on resistance to amoxicillin, cefotaxime, gentamicin, ofloxacin and cotrimoxazole.

Statistical Methods

We compared bacterial characteristics (1) in bacteremia from urinary versus digestive origin, (2) in children ≤3 versus >3 months of age, (3) in children versus adults presenting a community-acquired, urinary-source bacteremia. The cutoff of 3 months was chosen as the first 3 months of life correspond to a period of age with a known increased risk of E. coli bacteremia and meningitis related to the impaired innate immunity in this period.27,28 For the third comparison, all adults from the COLIBAFI cohort presenting a community-acquired, urinary-source bacteremia were included. Characteristics tested included phylogenetic group (from the results of the triplex PCR,18 grouped as B2 versus non-B2), virulence factors and PAIs, virulence and PAI scores, antimicrobial resistance, multidrug resistance and resistance score. Comparisons between groups were performed using nonparametric tests (Wilcoxon or Fisher exact tests).

We searched for risk factors associated with severity (yes/no). Analyses were performed in all children and in the ≤3 month of age subgroup. Clinical characteristics studied included gender, age (≤3 versus >3 months of age), immunodepression, nosocomial infection and portal of entry. Prematurity and birth weight were also studied in ≤3 months of age infants. The same bacterial characteristics than those described above were analyzed. Variables achieving P < 0.20 in univariate logistic regression analysis were entered into a multivariate logistic regression analysis to identify risk factors of severity. Using a forward selection method, we obtained a final model in which all risk factors had P < 0.05. First-order interaction was tested for significant variables. The model discrimination was assessed by the c-statistic and its 95% confidence interval (95% CI), and the model calibration was assessed by the Hosmer-Lemeshow goodness-of-fit test. Analyses were performed with SAS v9.3 (SAS Institute Inc., Cary, NC). All tests were 2-sided with a type-I error fixed to 0.05.


Clinical Characteristics

All 84 patients ≤18 years of age from the COLIBAFI cohort were included (Table 1). The most frequent portals of entry were the urinary tract (n = 51, 66.2%) and the digestive tract (n = 15, 19.5%). Eight children (10.4%) were immunocompromised: 4 had a history of hemopathy, 2 had had solid-organ transplantation, 1 had neutropenia and 1 had a congenital immunodeficiency. Fourteen children (16.7%) were transferred to ICU and 8 (9.5%) died during follow up. In total, 17 (20.2%) presented a severe bacteremia (95% CI: 12.2–30.4%).

Clinical and Bacteriological Characteristics of the 84 Children With E. coli Bacteremia

Bacterial Characteristics

Three children had a polymicrobial bacteremia. The most frequent E. coli phylogenetic group was B2 group (63.1%), followed by D, F and A groups (11.9%, 9.5% and 8.3%, respectively). No strain belonged to the E phylogenetic group. Among the B2 isolates, most represented clonal groups were the II (STc73) and IX (STc95) subgroups (37.7% for each group). Other clonal groups were I, IV, V, VI, VII and X groups (5.6%, 9.4%, 1.9%, 1.9%, 3.8% and 1.9%, respectively). All except 3 strains from the D phylogroup corresponded to the CGA. Three among the 7 strains of phylogroup A belonged to the STc10. Only 3 strains were not O-typed. Four O types were predominant: O6a (20.2%), O1 (19.0%), O2a and O2b (15.4%) and O7 (8.3%). Individual data for phylogenetic groups, clonal groups and O types are presented in Table (Supplemental Digital Content 1,

The proportion of extraintestinal virulence factors ranged from 6.0% for ibeA to 94.1% for fyuA and irp2 (see Table, Supplemental Digital Content 1, and Figure in Supplemental Digital Content 2, Likely, the frequency of PAIs ranged from 6.0% for the PAIgimA to 94.1% for the PAI IV536, with frequencies of 17.9%, 25.0%, 27.4%, 32.1% and 65.5% for PAI IIJ96, PAI ICFT073, PAI III536, PAIpks and PAIUSP, respectively. Median (min-max) virulence and PAI scores were 12 (1–9) and 2 (0–7), respectively. Only 1 strain belonging to the A1 clonal group (STc10) exhibited intraintestinal virulence genes, that is, afaD and aatA. Most isolates were resistant to amoxicillin (63.1%) and susceptible to ofloxacin and gentamicin (96.4% and 96.4%, Fig., Supplemental Digital Content 2, Three isolates were resistant to third-generation cephalosporins (3.6%) but none to imipenem, and 29 (34.5%) were resistant to cotrimoxazole. Three isolates (3.6%) were multidrug resistant.

Comparison of Urinary Versus Digestive Origin

The aerobactin gene iucC, the papGII allele and the PAI ICFT073 were more frequent in strains responsible of bacteremia from urinary origin than from digestive origin (P = 0.003, P = 0.03 and P = 0.05, respectively). There was no difference in phylogroup repartition or antimicrobial susceptibility according to portal of entry (Table 2).

Characteristics of the E. coli Isolates According to the Portal of Entry (Urinary- vs. Digestive-source Bacteremia)

Comparison of ≤3 Versus >3 Months of Age Children

Children ≤3 months of age (n = 43, 51%) were more frequently males than patients >3 months old (72.1% versus 41.5%, P < 0.01) and less frequently immunocompromised (0.0% versus 21.1%, P = 0.002). There was no significant difference found in portals of entry. Severe bacteremia were more frequently observed in children ≤3 months of age (32.6% versus 7.3%, P = 0.006). Thirteen children ≤3 months of age and 1 children >3 months of age were transferred to ICU, whereas 6 children ≤3 months of age and 2 children >3 months of age died during follow up. Bacterial characteristics are reported in Table 3. Virulence score was higher in ≤3 month of age children [median (min-max)= 15 (5–18) vs. 10 (1–19), P = 0.02], with 3 virulence factors being more frequently observed: neuC (K1 antigen, 55.8% vs. 26.8%, P = 0.009), irp2 (100% vs. 87.8%, P = 0.02) and fyuA (100% vs. 87.8%, P = 0.02). Antimicrobial resistance score was lower in ≤3 months of age infants [median (min-max) = 1 (0–4) vs. 1 (0–5), P = 0.01], with lower rates of resistance to amoxicillin (51.2% vs. 75.6%, P = 0.02) and cotrimoxazole (23.3% vs. 46.3%, P = 0.04).

Bacterial Characteristics in the ≤3 Versus >3 Month of Age Groups

Children Versus Adults Community-acquired, Urinary-source Bacteremia

Community-acquired urinary-source bacteremia concerned 45 children and 513 adults from the COLIBAFI cohort (Table 4). We observed more females (73.1% vs. 44.4%, P < 0.001) and a higher rate of immunodepression in adults than in children (23.9% vs. 4.9 %, P = 0.003). Severe bacteremia was more frequent in adults (11.9% vs. 0%, P = 0.01). There was no significant difference in phylogenetic groups (B2 vs. non-B2) nor in virulence scores, but 5 virulence factors were significantly more frequent in E. coli strains isolated from children (ie, iucC, iha, papC, papGII and sat), whereas hra was significantly less frequent. Adult isolates had a higher resistance score [median (min-max) = 1 (0–5) vs. 1 (0–4), P = 0.04].

Clinical and Bacterial Characteristics in Children and Adults With Community-acquired Urinary-source Bacteremia

Risk Factors of Severity

Results of the univariate and multivariate analyses to identify risk factors of severity in all the included children are reported Table 5. Because of some missing characteristics, the multivariate analysis was performed in 77 children among which 16 (20.8%) had a severe bacteremia. Non-urinary-source bacteremia [odds ratio (OR) = 0.01, 95% CI = 0.001–0.1] and being ≤3 months of age (OR = 7.7, 95% CI = 1.4–42.8) were the only risk factors of severity. The c-statistic of the final model was 0.93 (95% CI = 0.87–0.99). The P value of the Hosmer-Lemeshow test was 0.93, showing no model misspecification.

Risk Factors of Severity (Transfer in ICU or Death) From E. coli Bacteremia Identified by Univariate and Multivariate Analysis in all Included Children

Among the 43 children ≤3 months of age, the bacteremia was severe in 14 (32.6%). Birth weight was lower [median (min-max) = 1560 g (595–3580) versus 3285 g (910–5090), P < 0.01] and a preterm birth was more frequent (71.4% vs. 13.8%, P < 0.001) in ≤3 months of age children with severe bacteremia. Children ≤3 months of age with a severe bacteremia had a more frequently non-urinary (92.9% vs. 17.2%, P < 0.001) but digestive (50.0% vs. 10.3%, P < 0.01) portal of entry than those without. Because of some missing characteristics, the multivariate analysis was performed in 41 children among which 13 (31.7%) had a severe bacteremia. A non-urinary source of bacteremia was the only risk factor associated with severity (OR = 72.0, 95% CI = 7.2–796.9). Prematurity and birth weight were not found significantly associated with severity, nor any of the bacterial characteristics tested. The c-statistic for this final model was 0.89 (95% CI = 0.80–0.99). The P value of the Hosmer-Lemeshow test was >0.99, showing no model misspecification.


Few data are available about E. coli bacteremia in children despite its frequency and severity. To our knowledge, this is the first report combining bacterial and host characteristics in E. coli bacteremia in children.

We found a high proportion of strains belonging to the B2 and D phylogroups. Main clonal groups delineated from the B2 strains were the II and IX clonal groups, representing 23.8% each of all the strains. They correspond respectively to the STc73 and STc95.8,9 Strains belonging to STc131 (clonal group I) accounted for 3.6% in our study. We found a high proportion of CGA (STc69) among the D phylogroup strains, which accounted for 8.3% of all strains. All these CGA strains were resistant to cotrimoxazole. These results are similar to those obtained in adults. In a study of 300 uropathogenic E. coli, most common STc retrieved were STc73 (16.6%), STc131 (13.3%), STc69 (9%) and STc95 (6.3%).29 We also observed a predominance of 4 specific O types, that is O1, O2a/b, O6 and O7. These data indicate that some specific clones among the huge clonal diversity of the E. coli species4 are involved in extraintestinal infections. At the opposite, a high diversity in the pattern of presence or absence of virulence genes was evidenced, with 61 observed combinations (Table, Supplemental Digital Content 1, Altogether, these data support the fact that extraintestinal virulence in E. coli is because of multiple combinations of virulence genes6 arriving on specific genetic backgrounds.30 However, some specific virulence factors have been associated to specific syndromes. In their study of 123 children, Cheng et al16 showed that E. coli strains causing urinary-source bacteremia harbored more frequently papGI, iutA, traT, focG, afa, bmaE, kpsMT III, rfc and cvaL than strains causing acute pyelonephritis or acute lobar nephronia. Bonacorsi et al15 found that, in infants younger than 90 days, the presence of hly and/or iroN, as was that of hly and/or antigen K1, was associated in strains causing urinary tract infections with bacteremia. This group also found that, in young infants, papGII, sfa/foc and hly were more frequent in strains isolated from urosepsis than from meningitis.12,13 Similarly, in a study of 100 infants ≤3 months of age with bacteremia, these authors concluded that papGII and tcpC were more frequent, but ibeA less frequent, in E. coli originating from urinary tract infection than from gut translocation.17 In the same line, we found that iucC, which belongs to the aerobactin operon as iutA, and papGII were more frequent in urinary-source bacteremia (Table 2).

We took the opportunity of the large COLIBAFI cohort to compare community-acquired, urinary-source bacteremia between children and adults. We found that bacteremia in children were less severe, but involved strains exhibiting a specific repertoire of virulence genes and being more resistant than those isolated in adults. Children are usually considered to be more susceptible to infections because of an immature immune system. This apparently contra intuitive result might be explained by comorbidities in the adult group (23.9% of adults were immunocompromised,14 and only 4.9% of children). To our knowledge, no study compared the prevalence of resistant bacteria in children and in adults. Data of the European Antimicrobial Resistance Surveillance Network for the period 2005–2007 show that about 55% of E. coli strains were resistant to aminopenicillins.31 The higher prevalence of resistance observed in our cohort of children (63.1%) might be explained by the difference in antibiotic consumption among children. In France in 2004, the number of antimicrobial prescriptions for respiratory infections was almost 3 times higher in infants <2.5 years of age than in adults.32 This suggests that children have a higher selective pressure than adults, which could result in higher rates of infection with resistant bacteria.

In our study, we found a case-fatality rate of 9.5%, similar to that observed in a cohort of 2300 E. coli bacteremia in Canada.2 No bacterial characteristic was associated with severity defined as death or transfer to ICU during follow up. The severity outcome was driven by portal of entry and age. This observation was also made in adults, in whom we showed that host factors and the portal of entry outweighed bacterial determinants for predicting death from E. coli bacteremia.14 Houdouin et al27 examined patients’ characteristics and bacterial virulence mechanisms in lethal and nonlethal E. coli meningitis. In their retrospective study of 99 children ≤3 months of age, no pre-onset clinical factor was significantly associated with a fatal outcome. The only bacterial characteristic significantly associated with lethality was the absence of the aerobactin gene iucC. However, no multivariate analysis was performed.

Our study has several limitations. First, our prospective cohort included a rather small number of children. Bacteremia was severe only in 17 children, thus diminishing the power to determine risk factors of severity. However, we obtained a model with good predictive capacity (c-statistic of 0.93). Second, our study was performed during 2005–2007. Only 3 strains resistant to third-generation cephalosporins were isolated. In the recent years, the prevalence of resistant bacteria has increased. A study performed in Sweden in 2010 showed that 2.9% of healthy children were colonized with extended-spectrum beta-lactamase-producing Enterobacteriaceae. This figure rose to 8.4% in ill children.33 This could weaken our results, as bacteremia because of extended-spectrum beta-lactamase-producing E. coli have been associated with an increased risk of mortality in adults.34 Despite these limitations, our multicenter study with a prospective design is the first one focusing on E. coli bacteremia in children. Further studies should be performed on larger cohorts to confirm our findings.


We are grateful to Cécile Gateau for her technical help in the genotyping of the E. coli strains and to Jorge Blanco from the University of Santiago de Compostela, Lugo, Spain for serotyping of some strains.


1. Watson RS, Carcillo JA, Linde-Zwirble WT, et al. The epidemiology of severe sepsis in children in the United States. Am J Respir Crit Care Med. 2003;167:695–701
2. Laupland KB, Gregson DB, Church DL, et al. Incidence, risk factors and outcomes of Escherichia coli bloodstream infections in a large Canadian region. Clin Microbiol Infect. 2008;14:1041–1047
3. Doit C, Mariani-Kurkdjian P, Mahjoub-Messai F, et al. Epidemiology of pediatric community-acquired bloodstream infections in a children hospital in Paris, France, 2001 to 2008. Diagn Microbiol Infect Dis. 2010;66:332–335
4. Tenaillon O, Skurnik D, Picard B, et al. The population genetics of commensal Escherichia coli. Nat Rev Microbiol. 2010;8:207–217
5. Kaper JB, Nataro JP, Mobley HL. Pathogenic Escherichia coli. Nat Rev Microbiol. 2004;2:123–140
6. Touchon M, Hoede C, Tenaillon O, et al. Organised genome dynamics in the Escherichia coli species results in highly diverse adaptive paths. PLoS Genet. 2009;5:e1000344
7. Jaureguy F, Landraud L, Passet V, et al. Phylogenetic and genomic diversity of human bacteremic Escherichia coli strains. BMC Genomics.. 2008;9:560
8. Le Gall T, Clermont O, Gouriou S, et al. Extraintestinal virulence is a coincidental by-product of commensalism in B2 phylogenetic group Escherichia coli strains. Mol Biol Evol. 2007;24:2373–2384
9. Wirth T, Falush D, Lan R, et al. Sex and virulence in Escherichia coli: an evolutionary perspective. Mol Microbiol.. 2006;60:1136–1151
10. Moissenet D, Salauze B, Clermont O, et al. Meningitis caused by Escherichia coli producing TEM-52 extended-spectrum beta-lactamase within an extensive outbreak in a neonatal ward: epidemiological investigation and characterization of the strain. J Clin Microbiol. 2010;48:2459–2463
11. Johnson JR, Russo TA. Extraintestinal pathogenic Escherichia coli: “the other bad E. coli”. J Lab Clin Med. 2002;139:155–162
12. Nataro JP, Kaper JB. Diarrheagenic Escherichia coli. Clin Microbiol Rev. 1998;11:142–201
13. Bidet P, Mahjoub-Messai F, Blanco J, et al. Combined multilocus sequence typing and O serogrouping distinguishes Escherichia coli subtypes associated with infant urosepsis and/or meningitis. J Infect Dis. 2007;196:297–303
14. Lefort A, Panhard X, Clermont O, et al.COLIBAFI Group. Host factors and portal of entry outweigh bacterial determinants to predict the severity of Escherichia coli bacteremia. J Clin Microbiol. 2011;49:777–783
15. Bonacorsi S, Houdouin V, Mariani-Kurkdjian P, et al. Comparative prevalence of virulence factors in Escherichia coli causing urinary tract infection in male infants with and without bacteremia. J Clin Microbiol. 2006;44:1156–1158
16. Cheng CH, Tsau YK, Kuo CY, et al. Comparison of extended virulence genotypes for bacteria isolated from pediatric patients with urosepsis, acute pyelonephritis, and acute lobar nephronia. Pediatr Infect Dis J. 2010;29:736–740
17. Mahjoub-Messai F, Bidet P, Caro V, et al. Escherichia coli isolates causing bacteremia via gut translocation and urinary tract infection in young infants exhibit different virulence genotypes. J Infect Dis. 2011;203:1844–1849
18. Clermont O, Bonacorsi S, Bingen E. Rapid and simple determination of the Escherichia coli phylogenetic group. Appl Environ Microbiol. 2000;66:4555–4558
19. Clermont O, Christenson JK, Denamur E, et al. The Clermont Escherichia coli phylo-typing method revisited: improvement of specificity and detection of new phylo-groups. Environ Microbiol Rep. 2013;5:58–65
20. Johnson JR, Owens K, Manges AR, et al. Rapid and specific detection of Escherichia coli clonal group A by gene-specific PCR. J Clin Microbiol. 2004;42:2618–2622
21. Clermont O, Christenson J, Daubié A, et al. Development of an allele-specific PCR for Escherichia coli B2 sub-typing, a rapid and easy to perform substitute of multilocus sequence typing J Microbiol Meth.. 2014;;101::24–27
22. Clermont O, Johnson JR, Menard M, et al. Determination of Escherichia coli O types by allele-specific polymerase chain reaction: application to the O types involved in human septicemia. Diagn Microbiol Infect Dis. 2007;57:129–136
23. Clermont O, Olier M, Hoede C, et al. Animal and human pathogenic Escherichia coli strains share common genetic backgrounds. Infect Genet Evol. 2011;11:654–662
24. Johnson JR, Johnston B, Kuskowski MA, et al. Molecular epidemiology and phylogenetic distribution of the Escherichia coli pks genomic island. J Clin Microbiol. 2008;46:3906–3911
25. Hacker J, Blum-Oehler G, Mühldorfer I, et al. Pathogenicity islands of virulent bacteria: structure, function and impact on microbial evolution. Mol Microbiol.. 1997;23:1089–1097
26. Bingen-Bidois M, Clermont O, Bonacorsi S, et al. Phylogenetic analysis and prevalence of urosepsis strains of Escherichia coli bearing pathogenicity island-like domains. Infect Immun. 2002;70:3216–3226
27. Houdouin V, Bonacorsi S, Bidet P, et al. Association between mortality of Escherichia coli meningitis in young infants and non-virulent clonal groups of strains. Clin Microbiol Infect. 2008;14:685–690
28. Bachur R, Caputo GL. Bacteremia and meningitis among infants with urinary tract infections. Pediatr Emerg Care. 1995;11:280–284
29. Gibreel TM, Dodgson AR, Cheesbrough J, et al. Population structure, virulence potential and antibiotic susceptibility of uropathogenic Escherichia coli from Northwest England. J Antimicrob Chemother. 2012;67:346–356
30. Escobar-Páramo P, Clermont O, Blanc-Potard AB, et al. A specific genetic background is required for acquisition and expression of virulence factors in Escherichia coli. Mol Biol Evol. 2004;21:1085–1094
31. EARS-Net. . European Antimicrobial Resistance Surveillance Network - Annual Reports. 2007. Available at: Accessed on December 4, 2013.
32. Chahwakilian P, Huttner B, Schlemmer B, et al. Impact of the French campaign to reduce inappropriate ambulatory antibiotic use on the prescription and consultation rates for respiratory tract infections. J Antimicrob Chemother. 2011;66:2872–2879
33. Kaarme J, Molin Y, Olsen B, et al. Prevalence of extended-spectrum beta-lactamase-producing Enterobacteriaceae in healthy Swedish preschool children. Acta Paediatr. 2013;102:655–660
34. Rodríguez-Baño J, Picón E, Gijón P, et al.Spanish Network for Research in Infectious Diseases (REIPI). Community-onset bacteremia due to extended-spectrum beta-lactamase-producing Escherichia coli: risk factors and prognosis. Clin Infect Dis. 2010;50:40–48

bacteremia; Escherichia coli; children; portal of entry; prognosis; severity; death; intensive care units; virulence

Supplemental Digital Content

© 2014 by Lippincott Williams & Wilkins, Inc.